Classes#

class CategoricalProcess#

Categorical observation process.

Collection of univariate categorical emission distributions. In the multivariate case, use a collection of conditionally independent CategoricalProcesses observations[k] is a pointer to emission distribution for state k

Public Functions

void copy(const CategoricalProcess &process)#

Copy of a CategoricalProcess object.

Parameters:: process – [in] reference on a CategoricalProcess object.

void remove()#: Destruction of the data members of a CategoricalProcess object.

CategoricalProcess(int inb_state = 0, int inb_value = 0, bool observation_flag = false)#

Constructor of the CategoricalProcess class.

Parameters:

inb_state – [in] number of states,
inb_value – [in] number of categories,
observation_flag – [in] flag on the construction of the observation distributions.

CategoricalProcess(int inb_state, int inb_value, double **observation_probability)#

Constructor of the CategoricalProcess class.

Parameters:

inb_state – [in] number of states,
inb_value – [in] number of categories,
observation_probability – [in] pointer on the observation probabilities.

CategoricalProcess(int inb_state, Distribution **pobservation)#

Constructor of the CategoricalProcess class.

Parameters:

inb_state – [in] number of states,
pobservation – [in] pointer on the observation distributions.

~CategoricalProcess()#: Destructor of the CategoricalProcess class.

CategoricalProcess &operator=(const CategoricalProcess &process)#

Assignment operator of the CategoricalProcess class.

Parameters:: process – [in] reference on a CategoricalProcess object.
Returns:: CategoricalProcess object.

std::ostream &ascii_print(std::ostream &os, FrequencyDistribution **empirical_observation, FrequencyDistribution *marginal_distribution, bool exhaustive, bool file_flag, model_type model = HIDDEN_MARKOV) const#

Writing of a CategoricalProcess object.

Parameters:

os – [inout] stream,
empirical_observation – [in] pointer on the observation frequency distributions,
marginal_distribution – [in] pointer on the marginal frequency distribution,
exhaustive – [in] flag detail level,
file_flag – [in] flag file,
model – [in] model type.

std::ostream &spreadsheet_print(std::ostream &os, FrequencyDistribution **empirical_observation = NULL, FrequencyDistribution *marginal_distribution = NULL, model_type model = HIDDEN_MARKOV) const#

Writing of a CategoricalProcess object at the spreadsheet format.

Parameters:

os – [inout] stream,
empirical_observation – [in] pointer on the observation frequency distributions,
marginal_distribution – [in] pointer on the marginal frequency distribution,
model – [in] model type.

bool plot_print(const char *prefix, const char *title, int process, FrequencyDistribution **empirical_observation = NULL, FrequencyDistribution *marginal_distribution = NULL, model_type model = HIDDEN_MARKOV) const#

Plot of a CategoricalProcess object using Gnuplot.

Parameters:

prefix – [in] file prefix,
title – [in] figure title,
process – [in] observation process index,
empirical_observation – [in] pointer on the observation frequency distributions,
marginal_distribution – [in] pointer on the marginal frequency distribution,
model – [in] model type.

Returns:

error status.

void plotable_write(MultiPlotSet &plot, int &index, int process, FrequencyDistribution **empirical_observation = NULL, FrequencyDistribution *marginal_distribution = NULL, model_type model = HIDDEN_MARKOV) const#

Plot of a CategoricalProcess object.

Parameters:

plot – [in] reference on a MultiPlotSet object,
index – [in] MultiPlot index,
process – [in] observation process index,
empirical_observation – [in] pointer on the observation frequency distributions,
marginal_distribution – [in] pointer on the marginal frequency distribution,
model – [in] model type.

bool test_hidden() const#

test of the overlap between the categories observed in each state.

Returns:: hidden process or not.

void thresholding(double min_probability)#

Application of a threshold on the observation probabilities.

Parameters:: min_probability – [in] minimum probability.

void state_permutation(int *permut) const#

Permutation of observation distributions. The permutation validity should be checked by the calling function.

Parameters:: permut – [in] permutation.

int nb_parameter_computation(double min_probability) const#

Computation of the number of free parameters of a categorical observation process.

Parameters:: min_probability – [in] minimum probability.
Returns:: number of free parameters.

Distribution *mixture_computation(Distribution *pweight)#

Computation of a mixture of categorical observation distributions.

Parameters:: pweight – [in] pointer on the weight distribution.
Returns:: mixture of categorical observation distributions.

void init()#: Initialization of the observation probabilities.

Public Members

int nb_state#: number of states

int nb_value#: number of categories

Distribution **observation#: categorical observation distributions

Distribution *weight#: theoretical weights of observation distributions

Distribution *mixture#: mixture of observation distributions

Distribution *restoration_weight#: weights of observation distributions deduced from the restoration

Distribution *restoration_mixture#: mixture of observation distributions

Public Static Functions

static CategoricalProcess *parsing(StatError &error, ifstream &in_file, int &line, int nb_state, model_type model, bool hidden)#

Analysis of the format of categorical observation distributions.

Parameters:

error – [in] reference on a StatError object,
in_file – [in] stream,
line – [in] reference on the file line index,
nb_state – [in] number of states,
model – [in] model type,
hidden – [in] flag on the overlap of observation distributions.

Returns:

CategoricalProcess object.

static CategoricalProcess **old_parsing(StatError &error, ifstream &in_file, int &line, int nb_state, int &nb_output_process)#

Analysis of the format of observation distributions for a collection of categorical observation processes.

Parameters:

error – [in] reference on a StatError object,
in_file – [in] stream,
line – [in] reference on the file line index,
nb_state – [in] number of states,
nb_output_process – [in] number of observation processes.

Returns:

categorical observation processes.

class Chain#

Markov chain.

Public Functions

void parameter_copy(const Chain&)#

Copy of Markov chain parameters.

Parameters:: chain – [in] reference on a Chain object.

void copy(const Chain&)#

Copy of a Chain object.

Parameters:: chain – [in] reference on a Chain object.

void remove()#: Destruction of the data members of a Chain object.

Chain(process_type itype = ORDINARY, int inb_state = 0, bool init_flag = true)#

Constructor of the Chain class.

Parameters:

itype – [in] type,
inb_state – [in] number of states,
init_flag – [in] flag initialization.

Chain(process_type itype, int inb_state, int inb_row, bool init_flag)#

Constructor of the Chain class.

Parameters:

itype – [in] type,
inb_state – [in] number of states,
inb_row – [in] number of rows of the transition probability matrix,
init_flag – [in] flag initialization.

~Chain()#: Destructor of the Chain class.

Chain &operator=(const Chain &chain)#

Assignment operator of the Chain class.

Parameters:: chain – [in] reference on a Chain object.
Returns:: Chain object.

std::ostream &ascii_print(std::ostream &os, bool file_flag = false) const#

Writing of a Chain object.

Parameters:

os – [inout] stream,
file_flag – [in] file flag.

std::ostream &spreadsheet_print(std::ostream &os) const#

Writing of a Chain object at the spreadsheet format.

Parameters:: os – [inout] stream.

void create_cumul()#: Construction of the cumulative initial and transition distribution functions of a Chain object.

void cumul_computation()#: Computation of cumulative initial and transition distribution functions.

void remove_cumul()#: Destruction of the cumulative initial and transition distribution functions of a Chain object.

void log_computation()#: Log transform of Markov chain parameters.

bool **logic_transition_computation() const#

Construction of the matrix of possible transitions between states (adjacency matrix of the graph of possible transitions).

Returns:: matrix of possible transitions.

bool strongly_connected_component_research(StatError &error, bool **ilogic_transition = NULL) const#

Determination of the strongly connected components of a Markov chain.

Parameters:

error – [in] reference on a StatError object,
ilogic_transition – [in] matrix of possible transitions between states.

Returns:

error status.

void graph_accessibility_computation(bool **ilogic_transition)#

Computation of the accessibility of Markov chain states (graph-based approach).

Parameters:: ilogic_transition – [in] matrix of possible transitions between states.

void probability_accessibility_computation()#: Computation of the accessibility of Markov chain states (probabilistic approach).

void component_computation(bool **ilogic_transition = NULL)#

Extraction of the Markov chain classes (transient/recurrent/absorbing) from state accessibility.

Parameters:: ilogic_transition – [in] matrix of possible transitions between states.

bool parallel_initial_state() const#: Computation of the number of initial states in parallel (clustering structure).

void thresholding(double min_probability, bool semi_markov = false)#

Application of a threshold on the Markov chain parameters.

Parameters:

min_probability – [in] minimum probability,
semi_markov – [in] flag semi-Markov chain.

int nb_parameter_computation(double min_probability = 0.) const#

Computation of the number of free transition probabilities.

Parameters:: min_probability – [in] minimum probability.
Returns:: number of free transition probabilities.

double chi2_value_computation(const ChainData &chain_data) const#

Computation of the chi2 value for a Markov chain (goodness of fit test).

Parameters:: chain_data – [in] reference on a ChainData object.
Returns:: chi2 value.

void init(bool left_right, double self_transition)#

Initialization of Markov chain parameters.

Parameters:

left_right – [in] flag on the Markov chain structure,
self_transition – [in] self-transition probability.

double likelihood_computation(const ChainData &chain_data, bool initial_flag = true) const#

Computation of the log-likelihood of a Markov chain.

Parameters:

chain_data – [in] reference on a ChainData object,
initial_flag – [in] flag inclusion or not of the initial distribution in the log-likelihood computation.

Returns:

log-likelihood.

void chi2_fit(const ChainData &chain_data, Test &test) const#

Chi2 goodness of fit test for a Markov chain.

Parameters:

chain_data – [in] reference on a ChainData object,
test – [in] reference on a Test object.

int get_nb_state() const#: Return number of states.

Public Members

process_type type#: process type (ORDINARY/EQUILIBRIUM)

int nb_state#: number of states

int nb_row#: number of rows of the transition probability matrix

bool **accessibility#: state accessibility matrix

int nb_component#: number of classes

int *component_nb_state#: numbers of states per class

int **component#: classes

state_type *stype#: state types (TRANSIENT/RECURRENT/ABSORBING)

double *initial#: initial probabilities

double *cumul_initial#: cumulative initial distribution function

double **transition#: transition probability matrix

double **cumul_transition#: cumulative transition distribution functions

Public Static Functions

static Chain *parsing(StatError &error, ifstream &in_file, int &line, process_type type)#

Analysis of the format of a Chain object.

Parameters:

error – [in] reference on a StatError object,
in_file – [in] stream,
line – [in] reference on the file line index,
type – [in] process type (ORDINARY/EQUILIBRIUM).

Returns:

Chain object.

class ChainData : public stat_tool::ChainReestimation<int>#

Data structure corresponding to a Markov chain.

Public Functions

ChainData(const ChainData &chain_data)#

Constructor by copy of the ChainData class.

Parameters:: chain_data – [in] reference on a ChainData object.

int nb_parameter_computation() const#

Computation of the number of free transition probabilities.

Returns:: number of free transition probabilities.

void estimation(Chain &chain) const#

Estimation of first-order Markov chain parameters on the basis of initial state and transition frequencies.

Parameters:: chain – [in] reference on a Chain object.

template<typename Type> class ChainReestimation#

Data structure corresponding to a Markov chain.

Public Functions

void init()#: Initialization of a ChainReestimation object.

void copy(const ChainReestimation<Type> &chain_data)#

Copy of a ChainReestimation object.

Parameters:: chain_data – [in] reference on a ChainReestimation object.

void remove()#: Destruction of the data members of a ChainReestimation object.

ChainReestimation()#: Default constructor of a ChainReestimation object.

ChainReestimation(process_type itype, int inb_state, int inb_row, bool init_flag = false)#

Constructor of the ChainReestimation class.

Parameters:

itype – [in] type,
inb_state – [in] number of states,
inb_row – [in] number of rows of the transition probability matrix,
init_flag – [in] flag initialization.

~ChainReestimation()#: Destructor of the ChainReestimation class.

ChainReestimation<Type> &operator=(const ChainReestimation<Type> &chain_data)#

Assignment operator of the ChainReestimation class.

Parameters:: chain_data – [in] reference on a ChainReestimation object.
Returns:: ChainReestimation object.

std::ostream &print(std::ostream &os) const#

Writing of a ChainReestimation object.

Parameters:: os – [inout] stream.

Public Members

process_type type#: process type (ORDINARY/EQUILIBRIUM)

int nb_state#: number of states

int nb_row#: number of rows of the transition frequency matrix

Type *initial#: initial state frequencies

Type **transition#: transition frequency matrix

class Clusters : public stat_tool::DistanceMatrix #

Partitioning clustering results.

Public Functions

Clusters()#: Default constructor of the Clusters class.

Clusters(const DistanceMatrix &dist_matrix, int inb_cluster)#

Constructor of the Clusters class.

Parameters:

dist_matrix – [in] reference on a DistanceMatrix object,
inb_cluster – [in] number of clusters.

Clusters(const DistanceMatrix &dist_matrix, int inb_cluster, int *icluster_nb_pattern, int **cluster_pattern)#

Constructor of the Clusters class.

Parameters:

dist_matrix – [in] reference on a DistanceMatrix object,
inb_cluster – [in] number of clusters,
icluster_nb_pattern – [in] pointer on the cluster sizes,
cluster_pattern – [in] pointer on the cluster compositions.

~Clusters()#: Destructor of the Clusters class.

Clusters &operator=(const Clusters &clusters)#

Assignment operator of the Clusters class.

Parameters:: clusters – [in] reference on a Clusters object.
Returns:: Clusters object.

virtual std::ostream &line_write(std::ostream &os) const#

Writing on a line of a Clusters object.

Parameters:: os – [inout] stream.

virtual bool spreadsheet_write(StatError &error, const std::string path) const#

Writing of a Clusters object in a file at the spreadsheet format.

Parameters:

error – [in] reference on a StatError object,
path – [in] file path.

Returns:

error status.

virtual bool plot_write(StatError &error, const char *prefix, const char *title = NULL) const#

Plot of a Clusters object using Gnuplot.

Parameters:

error – [in] reference on a StatError object,
prefix – [in] file prefix,
title – [in] figure title.

Returns:

error status.

virtual MultiPlotSet *get_plotable() const#

Plot of a Clusters object.

Returns:: MultiPlotSet object.

void cluster_nb_pattern_computation()#: Computation of cluster sizes.

void pattern_distance_computation()#: Computation of individual-cluster distances and associated lengths on the basis of individual assignment to clusters.

void cluster_distance_computation_1()#: Computation of between-cluster distances and associated lengths.

void cluster_distance_computation_2()#: Computation of between-cluster distances and associated lengths on the basis of individual-cluster distances and associated lengths.

class Compound : public stat_tool::StatInterface, public stat_tool::Distribution #

Compound distribution.

Public Functions

Compound()#: Default constructor of the Compound class.

Compound(const DiscreteParametric &sum_dist, const DiscreteParametric &dist, double cumul_threshold = COMPOUND_THRESHOLD)#

Constructor of the Compound class.

Parameters:

sum_dist – [in] reference on the sum distribution,
dist – [in] reference on the basis distribution,
cumul_threshold – [in] threshold on the cumulative distribution function.

Compound(const DiscreteParametric &sum_dist, const DiscreteParametric &dist, compound_distribution type)#

Constructor of the Compound class.

Parameters:

sum_dist – [in] reference on the sum distribution,
dist – [in] reference on the basis distribution,
type – [in] unknown distribution type.

~Compound()#: Destructor of the Compound class.

Compound &operator=(const Compound &compound)#

Assignment operator of the Compound class.

Parameters:: compound – [in] reference on a Compound object.
Returns:: Compound object.

CompoundData *extract_data(StatError &error) const#

Extraction of the data part of a Compound object.

Parameters:: error – [in] reference on a StatError object.
Returns:: CompoundData object.

virtual std::ostream &line_write(std::ostream &os) const#

Writing on a single line of a Compound object.

Parameters:: os – [inout] stream.

virtual bool plot_write(StatError &error, const char *prefix, const char *title = NULL) const#

Plot of a Compound object using Gnuplot.

Parameters:

error – [in] reference on a StatError object,
prefix – [in] file prefix,
title – [in] figure title.

Returns:

error status.

virtual MultiPlotSet *get_plotable() const#

Plot of a Compound object.

Returns:: MultiPlotSet object.

void computation(int min_nb_value = 1, double cumul_threshold = COMPOUND_THRESHOLD, bool sum_flag = true, bool dist_flag = true)#

Computation of a compound distribution.

Parameters:

min_nb_value – [in] lower bound of the support,
cumul_threshold – [in] threshold on the cumulative distribution function,
sum_flag – [in] flag for the computation of the sum distribution,
dist_flag – [in] flag for the computation of the basis distribution.

CompoundData *simulation(StatError &error, int nb_element) const#

Simulation using a compound distribution.

Parameters:

error – [in] reference on a StatError object,
nb_element – [in] sample size.

Returns:

sample generated by a compound distribution.

Public Static Functions

static Compound *ascii_read(StatError &error, const std::string path, double cumul_threshold = COMPOUND_THRESHOLD)#

Construction of a Compound object from a file.

Parameters:

error – [in] reference on a StatError object,
path – [in] file path,
cumul_threshold – [in] threshold on the cumulative distribution function.

Returns:

Compound object.

class CompoundData : public stat_tool::StatInterface, public stat_tool::FrequencyDistribution #

Data structure corresponding to a compound distribution.

Public Functions

CompoundData()#: Default constructor of the CompoundData class.

CompoundData(const FrequencyDistribution &histo, const Compound &icompound)#

Constructor of the CompoundData class.

Parameters:

histo – [in] reference on a FrequencyDistribution object,
icompound – [in] reference on a Compound object.

CompoundData(const Compound &icompound)#

Constructor of the CompoundData class.

Parameters:: icompound – [in] reference on a Compound object.

~CompoundData()#: Destructor of the CompoundData class.

CompoundData &operator=(const CompoundData &compound_histo)#

Assignment operator of the CompoundData class.

Parameters:: compound_histo – [in] reference on a CompoundData object.
Returns:: CompoundData object.

DiscreteDistributionData *extract(StatError &error, compound_distribution type) const#

Extraction of the sum frequency distribution or of the basis frequency distribution.

Parameters:

error – [in] reference on a StatError object,
type – [in] frequency distribution type.

Returns:

DiscreteDistributionData object.

virtual std::ostream &line_write(std::ostream &os) const#

Writing on a single line of a CompoundData object.

Parameters:: os – [inout] stream.

virtual bool spreadsheet_write(StatError &error, const std::string path) const#

Writing of a CompoundData object in a file at the spreadsheet format.

Parameters:

error – [in] reference on a StatError object,
path – [in] file path.

Returns:

error status.

virtual bool plot_write(StatError &error, const char *prefix, const char *title = NULL) const#

Plot of a CompoundData object using Gnuplot.

Parameters:

error – [in] reference on a StatError object,
prefix – [in] file prefix,
title – [in] figure title.

Returns:

error status.

virtual MultiPlotSet *get_plotable() const#

Plot of a CompoundData object.

Returns:: MultiPlotSet object.

class ContinuousParametric#

Continuous parametric distribution.

Public Functions

void copy(const ContinuousParametric &dist)#

Copy of a ContinuousParametric object.

Parameters:: dist – [in] reference on a ContinuousParametric object.

ContinuousParametric(continuous_parametric iident = GAUSSIAN, double ilocation = D_INF, double idispersion = D_DEFAULT, double izero_probability = D_DEFAULT, angle_unit iunit = DEGREE)#

Constructor of the ContinuousParametric class.

Parameters:

iident – [in] identifier,
ilocation – [in] location parameter (Gaussian, inverse Gaussian, autoregressive model, von Mises), shape parameter (gamma, zero-inflated gamma) or intercept (linear model),
idispersion – [in] dispersion parameter (Gaussian, von Mises, linear model, autoregressive model) or scale parameter (gamma, zero-inflated gamma, inverse Gaussian),
izero_probability – [in] zero probability (zero-inflated gamma), slope (linear model) or autoregressive coefficient (autoregressive model),
iunit – [in] angle unit (von Mises).

~ContinuousParametric()#: Destructor of the ContinuousParametric class.

ContinuousParametric &operator=(const ContinuousParametric&)#

Assignment operator of the ContinuousParametric class.

Parameters:: dist – [in] reference on a ContinuousParametric object.
Returns:: ContinuousParametric object.

std::ostream &ascii_parameter_print(std::ostream &os, bool file_flag = false) const#

Writing of the parameters of a continuous distribution.

Parameters:

os – [inout] stream,
file_flag – [in] flag comment.

std::ostream &ascii_characteristic_print(std::ostream &os, bool file_flag = false) const#

Writing of the characteristics of a continuous distribution.

Parameters:

os – [inout] stream,
file_flag – [in] flag comment.

std::ostream &ascii_print(std::ostream &os, bool file_flag = false, bool cumul_flag = false, const Histogram *histo1 = NULL, const FrequencyDistribution *histo2 = NULL)#

Writing of a continuous distribution.

Parameters:

os – [inout] stream,
file_flag – [in] flag comment,
cumul_flag – [in] flag on the writing of the cumulative distribution function,
histo1 – [in] pointer on an Histogram object,
histo2 – [in] pointer on a FrequencyDistribution object.

std::ostream &spreadsheet_parameter_print(std::ostream &os) const#

Writing of the parameters of a continuous distribution at the spreadsheet format.

Parameters:: os – [inout] stream.

std::ostream &spreadsheet_print(std::ostream &os, bool cumul_flag = false, const Histogram *histo1 = NULL, const FrequencyDistribution *histo2 = NULL)#

Writing of a continuous distribution at the spreadsheet format.

Parameters:

os – [inout] stream,
cumul_flag – [in] flag on the writing of the cumulative distribution function,
histo1 – [in] pointer on an Histogram object,
histo2 – [in] pointer on a FrequencyDistribution object.

std::ostream &spreadsheet_characteristic_print(std::ostream &os) const#

Writing of the characteristics of a continuous distribution at the spreadsheet format.

Parameters:: os – [inout] stream.

std::ostream &plot_title_print(std::ostream &os) const#

Writing of the parameters of a continuous distribution at the Gnuplot format.

Parameters:: os – [inout] stream.

bool plot_print(const char *path, const Histogram *histo1 = NULL, const FrequencyDistribution *histo2 = NULL)#

Writing of a continuous distribution at the Gnuplot format.

Parameters:

path – [in] file path,
histo1 – [in] pointer on an Histogram object,
histo2 – [in] pointer on a FrequencyDistribution object.

Returns:

error status.

bool q_q_plot_print(const char *path, int nb_value, double **empirical_cdf) const#

Writing of a q-q plot at the Gnuplot format.

Parameters:

path – [in] file path,
nb_value – [in] number of values,
empirical_cdf – [in] pointer on the empirical cumulative distribution function.

Returns:

error status.

void plotable_write(SinglePlot &plot, const Histogram *histo1 = NULL, const FrequencyDistribution *histo2 = NULL)#

Plot of a continuous distribution.

Parameters:

plot – [in] reference on a SinglePlot object,
histo1 – [in] pointer on an Histogram object,
histo2 – [in] pointer on a FrequencyDistribution object.

void q_q_plotable_write(SinglePlot &plot, int nb_value, double **empirical_cdf) const#

Plot of a q-q plot.

Parameters:

plot – [in] reference on a SinglePlot object,
nb_value – [in] number of values,
empirical_cdf – [in] pointer on the empirical cumulative distribution function.

double **q_q_plot_computation(int nb_value, double **cdf) const#

Computation of a q-q plot.

Parameters:

nb_value – [in] number of values,
empirical_cdf – [in] pointer on the empirical cumulative distribution function.

Returns:

q-q plot.

int nb_parameter_computation() const#

Computation of the number of parameters of a continuous distribution.

Returns:: number of parameters.

double von_mises_mass_computation(double inf, double sup) const#

Computation of the density of a von Mises distribution integrated on an interval.

Parameters:

inf – [in] lower bound,
sup – [in] upper bound.

Returns:

integrated density.

double mass_computation(double inf, double sup) const#

Computation of the density of a distribution integrated on an interval.

Parameters:

inf – [in] lower bound,
sup – [in] upper bound.

Returns:

integrated density.

void von_mises_cumul_computation()#: Computation of the cumulative distribution function of a von Mises distribution.

double sup_norm_distance_computation(ContinuousParametric &dist)#

Computation of the distance between 2 continuous distributions (sup of the absolute difference between the cumulative distribution functions in the case of non-crossing cumulative distribution functions; in the general case, sum of sup on intervals between 2 crossings of cumulative distribution functions).

Parameters:: dist – [in] reference on a ContinuousParametric object.
Returns:: distance between 2 continuous distributions

double simulation()#

Simulation using a continuous distribution.

Returns:: generated value.

Public Members

continuous_parametric ident#: identifier

double shape#: shape parameter (gamma, zero-inflated gamma)

double location#: mean (Gaussian, inverse Gaussian, autoregressive model), mean direction (von Mises),

double intercept#: for linear model

double scale#: scale parameter (gamma, inverse Gaussian, zero-inflated gamma)

double dispersion#: standard deviation (Gaussian, linear model, autoregressive model), concentration (von Mises),

double zero_probability#: zero probability (zero-inflated gamma)

double slope#: for linear models

double autoregressive_coeff#: for autoregressive models

double min_value#: minimum value

double max_value#: maximum value

angle_unit unit#: unit (degree/radian - von Mises)

double slope_standard_deviation#: for linear models

double sample_size#: for linear and autoregressive models

double correlation#: for linear models

double determination_coeff#: for autoregressive models

double *cumul#: cumulative distribution function (von Mises)

Public Static Functions

static ContinuousParametric *parsing(StatError &error, std::ifstream &in_file, int &line, continuous_parametric last_ident)#

Analysis of the format of a ContinuousParametric object.

Parameters:

error – [in] reference on a StatError object,
in_file – [in] stream,
line – [in] reference on the file line index,
last_ident – [in] identifier of the last distribution in the list.

Returns:

ContinuousParametric object.

class ContinuousParametricProcess#

Continuous parametric observation process.

Collection of univariate continuous parametric emission distributions. In the multivariate case, use a collection of conditionally independent ContinuousParametricProcesses observations[k] is a pointer to emission distribution for state k

Public Functions

void copy(const ContinuousParametricProcess &process)#

Copy of a ContinuousParametricProcess object.

Parameters:: process – [in] reference on a ContinuousParametricProcess object.

void remove()#: Destruction of the data members of a ContinuousParametricProcess object.

ContinuousParametricProcess(int inb_state = 0)#

Constructor of the ContinuousParametricProcess class.

Parameters:: inb_state – [in] number of states.

ContinuousParametricProcess(int inb_state, ContinuousParametric **pobservation)#

Constructor of the ContinuousParametricProcess class.

Parameters:

inb_state – [in] number of states,
pobservation – [in] pointer on the observation distributions.

~ContinuousParametricProcess()#: Destructor of the ContinuousParametricProcess class.

ContinuousParametricProcess &operator=(const ContinuousParametricProcess &process)#

Assignment operator of the ContinuousParametricProcess class.

Parameters:: process – [in] reference on a ContinuousParametricProcess object.
Returns:: ContinuousParametricProcess object.

std::ostream &ascii_print(std::ostream &os, Histogram **observation_histogram, FrequencyDistribution **observation_distribution, Histogram *marginal_histogram, FrequencyDistribution *marginal_distribution, bool exhaustive, bool file_flag, model_type model = HIDDEN_MARKOV) const#

Writing of a ContinuousParametricProcess object.

Parameters:

os – [inout] stream,
observation_histogram – [in] pointer on the observation histograms,
observation_distribution – [in] pointer on the observation frequency distributions,
marginal_histogram – [in] pointer on the marginal histogram,
marginal_distribution – [in] pointer on the marginal frequency distribution,
exhaustive – [in] flag detail level,
file_flag – [in] flag file,
model – [in] model type.

std::ostream &spreadsheet_print(std::ostream &os, Histogram **observation_histogram = NULL, FrequencyDistribution **observation_distribution = NULL, Histogram *marginal_histogram = NULL, FrequencyDistribution *marginal_distribution = NULL, model_type model = HIDDEN_MARKOV) const#

Writing of a ContinuousParametricProcess object at the spreadsheet format.

Parameters:

os – [inout] stream,
observation_histogram – [in] pointer on the observation histograms,
observation_distribution – [in] pointer on the observation frequency distributions,
marginal_histogram – [in] pointer on the marginal histogram,
marginal_distribution – [in] pointer on the marginal frequency distribution,
model – [in] model type.

bool plot_print(const char *prefix, const char *title, int process, Histogram **observation_histogram = NULL, FrequencyDistribution **observation_distribution = NULL, Histogram *marginal_histogram = NULL, FrequencyDistribution *marginal_distribution = NULL, int nb_value = I_DEFAULT, double **empirical_cdf = NULL, model_type model = HIDDEN_MARKOV) const#

Plot of a ContinuousParametricProcess object at the Gnuplot format.

Parameters:

prefix – [in] file prefix,
title – [in] figure title,
process – [in] observation process index,
observation_histogram – [in] pointer on the observation histograms,
observation_distribution – [in] pointer on the observation frequency distributions,
marginal_histogram – [in] pointer on the marginal histogram,
marginal_distribution – [in] pointer on the marginal frequency distribution,
nb_value – [in] number of values,
empirical_cdf – [in] pointer on the empirical cumulative distribution function,
model – [in] model type.

Returns:

error status.

void plotable_write(MultiPlotSet &plot, int &index, int process, Histogram **observation_histogram = NULL, FrequencyDistribution **observation_distribution = NULL, Histogram *marginal_histogram = NULL, FrequencyDistribution *marginal_distribution = NULL, int nb_value = I_DEFAULT, double **empirical_cdf = NULL, model_type model = HIDDEN_MARKOV) const#

Plot of a ContinuousParametricProcess object.

Parameters:

plot – [in] reference on a MultiPlotSet object,
index – [in] MultiPlot index,
process – [in] observation process index,
observation_histogram – [in] pointer on the observation histograms,
observation_distribution – [in] pointer on the observation frequency distributions,
marginal_histogram – [in] pointer on the marginal histogram,
marginal_distribution – [in] pointer on the marginal frequency distribution,
nb_value – [in] number of values,
empirical_cdf – [in] pointer on the empirical cumulative distribution function,
model – [in] model type.

int nb_parameter_computation() const#

Computation of the number of free parameters of a continuous observation process.

Returns:: number of free parameters.

double mean_computation(Distribution *pweight) const#

Computation of the mean of a mixture of continuous observation distributions.

Parameters:: pweight – [in] pointer on the weight distribution.
Returns:: mixture mean.

double variance_computation(Distribution *pweight, double mean = D_INF) const#

Computation of the variance of a mixture of continuous observation distributions.

Parameters:

pweight – [in] pointer on weight distribution,
mean – [in] mean.

Returns:

mixture variance.

void select_unit(angle_unit iunit)#

Choice of the unit for von Mises observation distributions.

Parameters:: iunit – [in] unit (DEGREE/RADIAN).

void init(continuous_parametric iident, double min_value, double max_value, double mean, double variance)#

Initialization of continuous observation distributions.

Parameters:

iident – [in] distribution identifier,
min_value – [in] minimum value,
max_value – [in] maximum value,
mean – [in] empirical mean
variance – [in] empirical variance.

std::ostream &interval_computation(std::ostream &os)#

Computation of intervals on the basis of quantile and posterior probability criteria for a uniform weight distribution.

Parameters:: os – [inout] stream.

Public Members

int nb_state#: number of states

continuous_parametric ident#: identifiers of observation distributions

bool tied_location#: flag tied means (gamma, Gaussian)

bool tied_dispersion#: flag tied dispersion parameters (gamma, Gaussian, von Mises)

double offset#: offset for Gaussian mixture with evenly spaced means

angle_unit unit#: unit (degree/radian) for von Mises distributions

ContinuousParametric **observation#: continuous observation distributions

Distribution *weight#: theoretical weights of observation distributions

Distribution *restoration_weight#: weights of observation distributions deduced from the restoration

Public Static Functions

static ContinuousParametricProcess *parsing(StatError &error, ifstream &in_file, int &line, int nb_state, model_type model, continuous_parametric last_ident = VON_MISES)#

Analysis of the format of continuous observation distributions.

Parameters:

error – [in] reference on a StatError object,
in_file – [in] stream,
line – [in] reference on the file line index,
nb_state – [in] number of states,
model – [in] model type,
last_ident – [in] identifier of the last distribution in the list.

Returns:

ContinuousParametricProcess object.

class Convolution : public stat_tool::StatInterface, public stat_tool::Distribution #

Convolution of discrete distributions.

Public Functions

Convolution()#: Default constructor of the Convolution class.

Convolution(int nb_dist, const DiscreteParametric **pdist)#

Constructor of the Convolution class.

Parameters:

nb_dist – [in] number of distributions,
pdist – [in] pointer on the elementary distributions.

Convolution(const DiscreteParametric &known_dist, const DiscreteParametric &unknown_dist)#

Constructor of the Convolution class.

Parameters:

known_dist – [in] reference on the known distribution,
unknown_dist – [in] reference on the unknown distribution.

~Convolution()#: Destructor of the Convolution class.

Convolution &operator=(const Convolution &convol)#

Assignment operator of the Convolution class.

Parameters:: convol – [in] reference on a Convolution object.
Returns:: Convolution object.

DiscreteParametricModel *extract(StatError &error, int index) const#

Extraction of an elementary distribution.

Parameters:

error – [in] reference on a StatError object,
index – [in] distribution index.

Returns:

DiscreteParametricModel object.

ConvolutionData *extract_data(StatError &error) const#

Extraction of the data part of a Convolution object.

Parameters:: error – [in] reference on a StatError object.
Returns:: ConvolutionData object.

virtual std::ostream &line_write(std::ostream &os) const#

Writing on a single line of a Convolution object.

Parameters:: os – [inout] stream.

virtual bool plot_write(StatError &error, const char *prefix, const char *title = NULL) const#

Plot of a Convolution object using Gnuplot.

Parameters:

error – [in] reference on a StatError object,
prefix – [in] file prefix,
title – [in] figure title.

Returns:

error status.

virtual MultiPlotSet *get_plotable() const#

Plot of a Convolution object.

Returns:: MultiPlotSet object.

void computation(int min_nb_value = 1, double cumul_threshold = CONVOLUTION_THRESHOLD, bool *dist_flag = NULL)#

Computation of a convolution of discrete distributions.

Parameters:

min_nb_value – [in] upper bound of the elementary distribution supports,
cumul_threshold – [in] threshold on the cumulative distribution function,
dist_flag – [in] flags for computing elementary distributions.

ConvolutionData *simulation(StatError &error, int nb_element) const#

Simulation using a convolution of discrete distributions.

Parameters:

error – [in] reference on a StatError object,
nb_element – [in] sample size.

Returns:

ConvolutionData object.

Public Static Functions

static Convolution *build(StatError &error, int nb_dist, const DiscreteParametric **dist)#

Construction of a Convolution object on the basis of elementary distributions.

Parameters:

error – [in] reference on a StatError object,
nb_dist – [in] number of distributions,
dist – [in] pointer on the distributions.

Returns:

Convolution object.

static Convolution *ascii_read(StatError &error, const std::string path, double cumul_threshold = CONVOLUTION_THRESHOLD)#

Construction of a Convolution object from a file.

Parameters:

error – [in] reference on a StatError object,
path – [in] file path,
cumul_threshold – [in] threshold on the cumulative distribution function.

Returns:

Convolution object.

class ConvolutionData : public stat_tool::StatInterface, public stat_tool::FrequencyDistribution #

Data structure corresponding to a convolution of discrete distributions.

Public Functions

ConvolutionData()#: Default constructor of the ConvolutionData class.

ConvolutionData(const FrequencyDistribution &histo, int nb_dist)#

Constructor of the ConvolutionData class.

Parameters:

histo – [in] reference on a FrequencyDistribution object,
nb_dist – [in] number of frequency distributions.

ConvolutionData(const Convolution &convol)#

Constructor of the ConvolutionData class.

Parameters:: convol – [in] reference on a Convolution object.

~ConvolutionData()#: Destructor of the ConvolutionData class.

ConvolutionData &operator=(const ConvolutionData &convol_histo)#

Assignment operator of the ConvolutionData class.

Parameters:: convol_histo – [in] reference on a ConvolutionData object.
Returns:: ConvolutionData object.

DiscreteDistributionData *extract(StatError &error, int index) const#

Extraction of an elementary frequency distribution.

Parameters:

error – [in] reference on a StatError object,
index – [in] frequency distribution index.

Returns:

DiscreteDistributionData object.

virtual std::ostream &line_write(std::ostream &os) const#

Writing on a single line of a ConvolutionData object.

Parameters:: os – [inout] stream.

virtual bool spreadsheet_write(StatError &error, const std::string path) const#

Writing of a ConvolutionData object in a file at the spreadsheet format.

Parameters:

error – [in] reference on a StatError object,
path – [in] file path.

Returns:

error status.

virtual bool plot_write(StatError &error, const char *prefix, const char *title = NULL) const#

Plot of a ConvolutionData object using Gnuplot.

Parameters:

error – [in] reference on a StatError object,
prefix – [in] file prefix,
title – [in] figure title.

Returns:

error status.

virtual MultiPlotSet *get_plotable() const#

Plot of a ConvolutionData object.

Returns:: MultiPlotSet object.

class Curves#

Family of curves with frequencies.

Public Functions

void copy(const Curves&)#

Copy of a Curves object.

Parameters:: curves – [in] reference on a Curves object.

void smooth(const Curves &curves, int max_frequency)#

Copy of a Curves object with curve smoothing.

Parameters:

curves – [in] reference on a Curves object,
max_frequency – [in] threshold on the frequencies for the smoothing.

void remove()#: Destruction of data members of a Curves object.

Curves()#: Default constructor of the Curves class.

Curves(int inb_curve, int ilength, bool frequency_flag = false, bool index_parameter_flag = false, bool init_flag = true)#

Constructor of the Curves class.

Parameters:

inb_curve – [in] number of curves,
ilength – [in] curve length,
frequency_flag – [in] flag on the frequencies,
index_parameter_flag – [in] flag on the index parameters,
init_flag – [in] flag initialization.

Curves(const Curves &curves, curve_transformation transform = CURVE_COPY, int max_frequency = MAX_FREQUENCY)#

Constructor by copy of the Curves class.

Parameters:

curves – [in] reference on a Curves object,
transform – [in] type of transform (CURVE_COPY/SMOOTHING),
max_frequency – [in] threshold on the frequencies for the smoothing.

Curves(const Distribution &dist)#

Construction of a Curves object from a Distribution object.

Parameters:: dist – [in] reference on a Distribution object.

Curves(const FrequencyDistribution &histo)#

Construction of a Curves object from a FrequencyDistribution object.

Parameters:: histo – [in] reference on a FrequencyDistribution object.

~Curves()#: Destructor of the Curves class.

Curves &operator=(const Curves &curves)#

Assignment operator of the Curves class.

Parameters:: curves – [in] reference on a Curves object.
Returns:: Curves object.

std::ostream &ascii_print(std::ostream &os, bool file_flag = false, const Curves *curves = NULL) const#

Writing of 2 families of curves of the same length.

Parameters:

os – [inout] stream,
file_flag – [in] flag file,
curves – [in] pointer on a Curves object.

std::ostream &spreadsheet_print(std::ostream &os, const Curves *curves = NULL) const#

Writing of 2 families of curves of the same length at the spreadsheet format.

Parameters:

os – [inout] stream,
curves – [in] pointer on a Curves object.

int plot_length_computation() const#

Computation of the curve length to be plotted (Gnuplot output).

Returns:: curve length.

bool plot_print(const char *path, int ilength = I_DEFAULT, const Curves *curves_0 = NULL, const Curves *curves_1 = NULL) const#

Writing of 2 families of curves at the Gnuplot format.

Parameters:

path – [in] file path,
ilength – [in] curve length,
curves_0 – [in] pointer on a Curves object,
curves_1 – [in] pointer on a Curves object.

Returns:

error status.

bool plot_print_standard_residual(const char *path, double *standard_residual = NULL) const#

Writing of a curve and the corresponding standardized residuals at the Gnuplot format.

Parameters:

path – [in] file path,
standard_residual – [in] pointer on the standardized residuals.

Returns:

error status.

void plotable_write(int index, SinglePlot &plot) const#

Writing of a curve.

Parameters:

index – [in] curve index,
plot – [in] reference on a SinglePlot object.

void plotable_write(MultiPlot &plot) const#

Writing of the curve family.

Parameters:: plot – [in] reference on a MultiPlot object.

void plotable_frequency_write(SinglePlot &plot) const#

Writing of frequencies.

Parameters:: plot – [in] reference on a SinglePlot object.

int max_frequency_computation() const#

Computation of the maximum frequency of a Curves object.

Returns:: maximum frequency.

int nb_element_computation() const#

Computation of the cumulative frequency of a Curves object.

Returns:: cumulative frequency.

double mean_computation(int index) const#

Computation of the curve mean.

Parameters:: index – [in] curve index.
Returns:: mean.

double total_square_sum_computation(int index, double mean) const#

Computation of the total variation for a curve.

Parameters:

index – [in] curve index,
mean – [in] mean.

Returns:

total variation.

Public Members

int nb_curve#: number of curves

int length#: curve length

int offset#: index of the beginning of the curves

int *index_parameter#: explicit index parameters

int *frequency#: frequency for each index

double **point#: curves

class Dendrogram : public stat_tool::StatInterface #

Hierarchical clustering results.

Public Functions

Dendrogram()#: Default constructor of the Dendrogram class.

Dendrogram(const DistanceMatrix &dist_matrix, cluster_scale iscale)#

Constructor of the Dendrogram class.

Parameters:

dist_matrix – [in] reference on a DistanceMatrix object,
iscale – [in] cluster distance scale.

~Dendrogram()#: Destructor of the Dendrogram class.

Dendrogram &operator=(const Dendrogram &dendrogram)#

Assignment operator of the Dendrogram class.

Parameters:: dendrogram – [in] reference on a Dendrogram object.
Returns:: Dendrogram object.

virtual std::ostream &line_write(std::ostream &os) const#

Writing on a single line of a Dendrogram object.

Parameters:: os – [inout] stream.

virtual bool spreadsheet_write(StatError &error, const std::string path) const#

Writing of a Dendrogram object in a file at the spreadsheet format.

Parameters:

error – [in] reference on a StatError object,
path – [in] file path.

Returns:

error status.

class DiscreteDistributionData : public stat_tool::StatInterface, public stat_tool::FrequencyDistribution #

Public Functions

DiscreteDistributionData(const FrequencyDistribution &histo, const Distribution *dist)#

Construction of a DiscreteDistributionData object from a FrequencyDistribution and a Distribution objects.

Parameters:

histo – [in] reference on a FrequencyDistribution object,
dist – [in] pointer on a Distribution object.

DiscreteDistributionData(const FrequencyDistribution &histo, const DiscreteParametric *dist)#

Construction of a DiscreteDistributionData object from a FrequencyDistribution and a DiscreteParametric objects.

Parameters:

histo – [in] reference on a FrequencyDistribution object,
dist – [in] pointer on a DiscreteParametric object.

DiscreteDistributionData(const DiscreteDistributionData &histo, bool model_flag = true)#

Constructor by copy of the DiscreteDistributionData class.

Parameters:

histo – [in] reference on a DiscreteDistributionData object,
model_flag – [in] flag copy of the included DiscreteParametricModel object.

~DiscreteDistributionData()#: Destructor of the DiscreteDistributionData class.

DiscreteDistributionData &operator=(const DiscreteDistributionData &histo)#

Assignment operator of the DiscreteDistributionData class.

Parameters:: histo – [in] reference on a DiscreteDistributionData object.
Returns:: DiscreteDistributionData object.

DiscreteParametricModel *extract_model(StatError &error) const#

Extraction of the DiscreteParametricModel object included in a DiscreteDistributionData object.

Parameters:: error – [in] reference on a StatError object.
Returns:: DiscreteParametricModel object.

virtual std::ostream &line_write(std::ostream &os) const#

Writing on a single line of a DiscreteDistributionData object.

Parameters:: os – [inout] stream.

virtual bool spreadsheet_write(StatError &error, const std::string path) const#

Writing of a DiscreteDistributionData object in a file at the spreadsheet format.

Parameters:

error – [in] reference on a StatError object,
path – [in] file path.

Returns:

error status.

virtual bool plot_write(StatError &error, const char *prefix, const char *title = NULL) const#

Plot of a DiscreteDistributionData object using Gnuplot.

Parameters:

error – [in] reference on a StatError object,
prefix – [in] file prefix,
title – [in] figure title.

Returns:

error status.

virtual MultiPlotSet *get_plotable() const#

Plot of a DiscreteDistributionData object.

Returns:: MultiPlotSet object.

Public Static Functions

static DiscreteDistributionData *ascii_read(StatError &error, const std::string path)#

Construction of a FrequencyDistribution object from a file. Format: n rows with ordered value and associated frequency.

Parameters:

error – [in] reference on a StatError object,
path – [in] file path.

class DiscreteMixture : public stat_tool::StatInterface, public stat_tool::Distribution #

Mixture of discrete distributions.

Public Functions

DiscreteMixture()#: Default constructor of the DiscreteMixture class.

DiscreteMixture(int inb_component, double *pweight, const DiscreteParametric **pcomponent)#

Constructor of the DiscreteMixture class.

Parameters:

inb_component – [in] number of components,
iweight – [in] weights,
pcomponent – [in] pointer on the components.

DiscreteMixture(const DiscreteMixture &mixt, bool *component_flag, int inb_value)#

Constructor of the DiscreteMixture class.

Parameters:

mixt – [in] reference on a DiscreteMixture object,
component_flag – [in] flags on the components to be copied,
inb_value – [in] upper bound of the support.

DiscreteMixture(int inb_component, const DiscreteParametric **pcomponent)#

Constructor of the DiscreteMixture class.

Parameters:

inb_component – [in] number of components,
pcomponent – [in] pointer on the components.

~DiscreteMixture()#: Destructor of the DiscreteMixture class.

DiscreteMixture &operator=(const DiscreteMixture &mixt)#

Assignment operator of the DiscreteMixture class.

Parameters:: mixt – [in] reference on a DiscreteMixture object.
Returns:: DiscreteMixture object.

DiscreteParametricModel *extract(StatError &error, int index) const#

Extraction of a component.

Parameters:

error – [in] reference on a StatError object,
index – [in] component index.

Returns:

DiscreteParametricModel object.

DiscreteMixtureData *extract_data(StatError &error) const#

Extraction of the DiscreteMixtureData object included in a DiscreteMixture object.

Parameters:: error – [in] reference on a StatError object.
Returns:: DiscreteMixtureData object.

virtual std::ostream &line_write(std::ostream &os) const#

Writing on a single line of a DiscreteMixture object.

Parameters:: os – [inout] stream.

virtual bool plot_write(StatError &error, const char *prefix, const char *title = NULL) const#

Plot of a DiscreteMixture object using Gnuplot.

Parameters:

error – [in] reference on a StatError object,
prefix – [in] file prefix,
title – [in] figures title.

Returns:

error status.

virtual MultiPlotSet *get_plotable() const#

Plot of a DiscreteMixture object.

Returns:: MultiPlotSet object.

void computation(int min_nb_value = 1, double cumul_threshold = CUMUL_THRESHOLD, bool component_flag = true)#

Computation of a mixture of discrete distributions.

Parameters:

min_nb_value – [in] lower bound of the component support,
cumul_threshold – [in] threshold on the cumulative distribution function,
component_flag – [in] flag for the component computation.

double likelihood_computation(const DiscreteMixtureData &mixt_histo) const#

Computation of the log-likelihood of a mixture of discrete distributions for a DiscreteMixtureData object.

Parameters:: mixt_histo – [in] reference on a DiscreteMixtureData object.
Returns:: log-likelihood.

DiscreteMixtureData *simulation(StatError &error, int nb_element) const#

Simulation using a mixture of discrete distributions.

Parameters:

error – [in] reference on a StatError object,
nb_element – [in] sample size.

Returns:

DiscreteMixtureData object.

Public Static Functions

static DiscreteMixture *build(StatError &error, int nb_component, double *weight, const DiscreteParametric **component)#

Construction of a DiscreteMixture object on the basis of weights and components.

Parameters:

error – [in] reference on a StatError object,
nb_component – [in] number of components,
weight – [in] component weights,
component – [in] pointer on the components.

Returns:

DiscreteMixture object.

static DiscreteMixture *ascii_read(StatError &error, const std::string path, double cumul_threshold = CUMUL_THRESHOLD)#

Construction of a DiscreteMixture object from a file.

Parameters:

error – [in] reference on a StatError object,
path – [in] file path,
cumul_threshold – [in] threshold on the cumulative distribution function.

Returns:

DiscreteMixture object.

class DiscreteMixtureData : public stat_tool::StatInterface, public stat_tool::FrequencyDistribution #

Data structure corresponding to a mixture of discrete distributions.

Public Functions

DiscreteMixtureData()#: Default constructor of the DiscreteMixtureData class.

DiscreteMixtureData(const FrequencyDistribution &histo, int inb_component)#

Constructor of the DiscreteMixtureData class.

Parameters:

histo – [in] reference on a FrequencyDistribution object,
inb_component – [in] number of components.

DiscreteMixtureData(const FrequencyDistribution &histo, const DiscreteMixture *pmixture)#

Constructor of the DiscreteMixtureData class.

Parameters:

histo – [in] reference on a FrequencyDistribution object,
pmixture – [in] pointer on a DiscreteMixture object.

DiscreteMixtureData(const DiscreteMixture &mixt)#

Constructor of the DiscreteMixtureData class.

Parameters:: mixt – [in] reference on a DiscreteMixture object.

~DiscreteMixtureData()#: Destructor of the DiscreteMixtureData class.

DiscreteMixtureData &operator=(const DiscreteMixtureData &mixt_histo)#

Assignment operator of the DiscreteMixtureData class.

Parameters:: mixt_histo – [in] reference on a DiscreteMixtureData object.
Returns:: DiscreteMixtureData object.

DiscreteDistributionData *extract(StatError &error, int index) const#

Extraction of a component frequency distribution.

Parameters:

error – [in] reference on a StatError object,
index – [in] component index.

Returns:

DiscreteDistributionData object.

virtual std::ostream &line_write(std::ostream &os) const#

Writing on a single line of a DiscreteMixtureData object.

Parameters:: os – [inout] stream.

virtual bool spreadsheet_write(StatError &error, const std::string path) const#

Writing of a DiscreteMixtureData object in a file at the spreadsheet format.

Parameters:

error – [in] reference on a StatError object,
path – [in] file path.

Returns:

error status.

virtual bool plot_write(StatError &error, const char *prefix, const char *title = NULL) const#

Plot of a DiscreteMixtureData object using Gnuplot.

Parameters:

error – [in] reference on a StatError object,
prefix – [in] file prefix,
title – [in] figure title.

Returns:

error status.

virtual MultiPlotSet *get_plotable() const#

Plot of a DiscreteMixtureData object.

Returns:: MultiPlotSet object.

double information_computation() const#

Computation of the information quantity of a DiscreteMixtureData object.

Returns:: information quantity.

class DiscreteParametric : public stat_tool::Distribution #

Discrete parametric distribution.

Subclassed by stat_tool::DiscreteParametricModel, stat_tool::Forward

Public Functions

void init(int iinf_bound, int isup_bound, double iparameter, double iprobability)#

Initialization of the parameters of a discrete parametric distribution.

Parameters:

iinf_bound – [in] lower bound,
isup_bound – [in] upper bound (binomial, uniform),
iparameter – [in] parameter (Poisson, negative binomial, Poisson geometric),
iprobability – [in] probability (binomial, negative binomial, Poisson geometric).

void init(discrete_parametric iident, int iinf_bound, int isup_bound, double iparameter, double iprobability)#

Initialization of the identifier and parameters of a discrete parametric distribution.

Parameters:

iident – [in] identifier,
iinf_bound – [in] lower bound,
isup_bound – [in] upper bound (binomial, uniform),
iparameter – [in] parameter (Poisson, negative binomial, Poisson geometric),
iprobability – [in] probability (binomial, negative binomial, Poisson geometric).

void copy(const DiscreteParametric &dist)#

Copy of a DiscreteParametric object.

Parameters:: dist – [in] reference on a DiscreteParametric object.

DiscreteParametric(int inb_value = 0, discrete_parametric iident = CATEGORICAL, int iinf_bound = I_DEFAULT, int isup_bound = I_DEFAULT, double iparameter = D_DEFAULT, double iprobability = D_DEFAULT)#

Constructor of the DiscreteParametric class.

Parameters:

inb_value – [in] number of values,
iident – [in] identifier,
iinf_bound – [in] lower bound,
isup_bound – [in] upper bound (binomial, uniform),
iparameter – [in] parameter (Poisson, negative binomial, Poisson geometric),
iprobability – [in] probability (binomial, negative binomial, Poisson geometric).

DiscreteParametric(discrete_parametric iident, int iinf_bound, int isup_bound, double iparameter, double iprobability, double cumul_threshold = CUMUL_THRESHOLD)#

Constructor of the DiscreteParametric class.

Parameters:

iident – [in] identifier,
iinf_bound – [in] lower bound,
isup_bound – [in] upper bound (binomial, uniform),
iparameter – [in] parameter (Poisson, negative binomial, Poisson geometric),
iprobability – [in] probability (binomial, negative binomial, Poisson geometric),
cumul_threshold – [in] threshold on the cumulative distribution function.

DiscreteParametric(int iinf_bound, int ino_segment, int isequence_length)#

Constructor of the DiscreteParametric class (prior segment length distribution).

Parameters:

iinf_bound – [in] lower bound,
ino_segment – [in] number of segments,
isequence_length – [in] sequence length.

DiscreteParametric(const Distribution &dist, int ialloc_nb_value = I_DEFAULT)#

Construction of a DiscreteParametric object from a Distribution object.

Parameters:

dist – [in] reference on a Distribution object,
ialloc_nb_value – [in] number of allocated values.

DiscreteParametric(const Distribution &dist, double scaling_coeff)#

Construction of a DiscreteParametric object from a Distribution object applying a scaling operation.

Parameters:

dist – [in] reference on a Distribution object,
scaling_coeff – [in] scaling factor.

DiscreteParametric(const DiscreteParametric &dist, double scaling_coeff)#

Construction of an inter-event distribution from an initial inter-event distribution up- or down-scaling the time scale.

Parameters:

dist – [in] reference on an inter-event distribution,
scaling_coeff – [in] scaling factor.

DiscreteParametric(const FrequencyDistribution &histo)#

Construction of a DiscreteParametric object from a FrequencyDistribution object.

Parameters:: histo – [in] reference on a FrequencyDistribution object.

DiscreteParametric(const DiscreteParametric &dist, distribution_transformation transform = DISTRIBUTION_COPY, int ialloc_nb_value = I_DEFAULT)#

Constructor by copy of the DiscreteParametric class.

Parameters:

dist – [in] reference on a DiscreteParametric object,
transform – [in] type of transform (DISTRIBUTION_COPY/NORMALIZATION),
ialloc_nb_value – [in] number of allocated values.

DiscreteParametric &operator=(const DiscreteParametric &dist)#

Assignment operator of the DiscreteParametric class.

Parameters:: dist – [in] reference on a DiscreteParametric object.
Returns:: DiscreteParametric object.

std::ostream &ascii_print(std::ostream &os) const#

Writing of the parameters of a discrete distribution.

Parameters:: os – [inout] stream.

std::ostream &ascii_parametric_characteristic_print(std::ostream &os, bool shape = false, bool comment_flag = false) const#

Writing of the characteristics of a discrete parametric distribution.

Parameters:

os – [inout] stream,
shape – [in] flag on the writing of the shape characteristics,
comment_flag – [in] flag comment.

std::ostream &spreadsheet_print(std::ostream &os) const#

Writing of the parameters of a discrete distribution at the spreadsheet format.

Parameters:: os – [inout] stream.

std::ostream &spreadsheet_parametric_characteristic_print(std::ostream &os, bool shape = false) const#

Writing of the characteristics of a discrete parametric distribution at the spreadsheet format.

Parameters:

os – [inout] stream,
shape – [in] flag on the writing of the shape characteristics.

virtual std::ostream &plot_title_print(std::ostream &os) const#

Writing of the parameters of a discrete parametric distribution at the Gnuplot format.

Parameters:: os – [inout] stream.

int nb_parameter_computation()#

Computation of the number of parameters of a discrete parametric distribution.

Returns:: number of parameters.

void nb_parameter_update()#: Update of the number of parameters of a discrete parametric distribution.

double parametric_mean_computation() const#

Computation of the theooretical mean of a discrete parametric distribution.

Returns:: theoretical mean

double parametric_variance_computation() const#

Computation of the theoretical variance of a discrete parametric distribution.

Returns:: theoretical variance.

double parametric_skewness_computation() const#

Computation of the theoretical coefficient of skewness of a discrete parametric distribution.

Returns:: theoretical coefficient of skewness.

double parametric_kurtosis_computation() const#

Computation of the theoretical excess kurtosis of a discrete parametric distribution: excess kurtosis = coefficient of kurtosis - 3.

Returns:: theoretical excess kurtosis

double sup_norm_distance_computation(const DiscreteParametric &dist) const#

Computation of the distance between 2 discrete parametric distributions (sup of the absolute difference between the cumulative distribution functions in the case of non-crossing cumulative distribution functions; in the general case, sum of sup on intervals between 2 crossings of cumulative distribution functions).

Parameters:: dist – [in] reference on a DiscreteParametric object.
Returns:: distance between 2 discrete parametric distributions

void binomial_computation(int inb_value, distribution_computation mode)#

Computation of the probability mass function of a binomial distribution.

Parameters:

inb_value – [in] number of values,
mode – [in] computation mode (STANDARD/RENEWAL).

void poisson_computation(int inb_value, double cumul_threshold, distribution_computation mode)#

Computation of the probability mass function of a Poisson distribution. The number of values is determined using a threshold on the cumulative distribution function or using a predefined bound.

Parameters:

inb_value – [in] number of values,
cumul_threshold – [in] threshold on the cumulative distribution function,
mode – [in] computation mode (STANDARD/RENEWAL).

void negative_binomial_computation(int inb_value, double cumul_threshold, distribution_computation mode)#

Computation of the probability mass function of a negative binomial distribution. The number of values is determined using a threshold on the cumulative distribution function or using a predefined bound.

Parameters:

inb_value – [in] number of values,
cumul_threshold – [in] threshold on the cumulative distribution function,
mode – [in] computation mode (STANDARD/RENEWAL).

void geometric_poisson_computation(int inb_value, double cumul_threshold)#

Computation of the probability mass function of a Poisson geometric distribution. The number of values is determined using a threshold on the cumulative distribution function or using a predefined bound.

Parameters:

inb_value – [in] number of values,
cumul_threshold – [in] threshold on the cumulative distribution function.

void uniform_computation()#: Computation of the probability mass function of a discrete uniform distribution.

void prior_segment_length_computation()#: Computation of the prior segment length distribution corresponding to the assumption of a uniform prior distribution for the possible segmentations.

void computation(int min_nb_value = 1, double cumul_threshold = CUMUL_THRESHOLD)#

Computation of the probability mass function of a parametric discrete distribution (binomial, Poisson, negative binomial, uniform compound Poisson geometric, prior segment length distribution for a multiple change-point model).

Parameters:

min_nb_value – [in] minimum number of values,
cumul_threshold – [in] threshold on the cumulative distribution function.

int simulation() const#

Simulation using the rejection sampling method. Principle: A point (x, Px) is drawn in the rectangle [xmin, xmax] x [0. ,Pmax]. If the point is below the distribution, the x value is kept; If not, a new point is drawn.

Returns:: generated value.

double renewal_likelihood_computation(const Forward &forward_dist, const FrequencyDistribution &within, const FrequencyDistribution &backward, const FrequencyDistribution &forward, const FrequencyDistribution *no_event) const#

Computation of the log-likelihood of a discrete-time renewal process for time interval data.

Parameters:

forward_dist – [in] forward recurrence time distribution,
within – [in] complete time interval frequency distribution,
backward – [in] backward recurrence time frequency distribution,
forward – [in] forward recurrence time frequency distribution,
no_event – [in] observation period frequency distribution for the case of no event.

Returns:

log-likelihood.

void expectation_step(const FrequencyDistribution &within, const FrequencyDistribution &backward, const FrequencyDistribution &forward, const FrequencyDistribution *no_event, Reestimation<double> *inter_event_reestim, Reestimation<double> *length_bias_reestim, int iter) const#

Computation of the reestimation quantities corresponding to the inter-event distribution (EM algorithm of an equilibrium renewal process estimated on the basis of time interval data).

Parameters:

within – [in] complete time interval frequency distribution,
backward – [in] backward recurrence time frequency distribution,
forward – [in] forward recurrence time frequency distribution,
no_event – [in] observation period frequency distribution for the case of no event,
inter_event_reestim – [in] pointer on the reestimation quantities of the inter-event distribution
length_bias_reestim – [in] pointer on the reestimation quantities of the length-biased distribution,
iter – [in] EM iteration.

double state_occupancy_likelihood_computation(const FrequencyDistribution &sojourn_time, const FrequencyDistribution &final_run) const#

Computation of the log-likelihood of a state occupancy distribution of an ordinary semi-Markov chain.

Parameters:

sojourn_time – [in] frequency distribution of complete sojourn times,
final_run – [in] frequency distribution of right-censored sojourn times.

Returns:

log-likelihood of the estimated distribution.

double state_occupancy_likelihood_computation(const Forward &forward, const FrequencyDistribution &sojourn_time, const FrequencyDistribution &initial_run, const FrequencyDistribution &final_run, const FrequencyDistribution &single_run) const#

Computation of the log-likelihood of a state occupancy distribution of an equilibrium semi-Markov chain.

Parameters:

forward – [in] forward sojourn time distribution,
sojourn_time – [in] frequency distribution of complete sojourn times,
initial_run – [in] frequency distribution of left-censored sojourn times,
final_run – [in] frequency distribution of right-censored sojourn times,
single_run – [in] frequency distribution of sequence lengths for the case of a single visited state.

Returns:

log-likelihood of the estimated distribution.

void expectation_step(const FrequencyDistribution &sojourn_time, const FrequencyDistribution &final_run, Reestimation<double> *occupancy_reestim, int iter) const#

Computation of the reestimation quantities of a state occupancy distribution (EM estimator of an ordinary semi-Markov chain).

Parameters:

sojourn_time – [in] frequency distribution of complete sojourn times,
final_run – [in] frequency distribution of right-censored sojourn times,
occupancy_reestim – [in] pointer on the reestimation quantities of the state occupancy distribution,
iter – [in] EM iteration.

void expectation_step(const FrequencyDistribution &sojourn_time, const FrequencyDistribution &initial_run, const FrequencyDistribution &final_run, const FrequencyDistribution &single_run, Reestimation<double> *occupancy_reestim, Reestimation<double> *length_bias_reestim, int iter, bool combination = false, duration_distribution_mean_estimator mean_estimator = COMPUTED) const#

Computation of the reestimation quantities of a state occupancy distribution (EM estimator of an equilibrium semi-Markov chain).

Parameters:

sojourn_time – [in] frequency distribution of complete sojourn times,
initial_run – [in] frequency distribution of left-censored sojourn times,
final_run – [in] frequency distribution of right-censored sojourn times,
single_run – [in] frequency distribution of sequence lengths for the case of a single visited state,
occupancy_reestim – [in] pointer on the reestimation quantities of the state occupancy distribution,
length_bias_reestim – [in] pointer on the reestimation quantities of the length-biased distribution,
iter – [in] EM iteration,
combination – [in] combination or not of the reestimation quantities,
mean_estimator – [in] method for the computation of the mean of the state occupancy distribution.

Public Members

discrete_parametric ident#: identifier

int inf_bound#: lower bound

int sup_bound#: upper bound (binomial, uniform)

int no_segment#: number of segments (prior segment length distribution)

double parameter#: parameter (Poisson, negative binomial, Poisson geometric)

double probability#: probability of success (binomial, negative binomial, Poisson geometric)

int sequence_length#: sequence length (prior segment length distribution)

Public Static Functions

static DiscreteParametric *parsing(StatError &error, std::ifstream &in_file, int &line, discrete_parametric last_ident = NEGATIVE_BINOMIAL, double cumul_threshold = CUMUL_THRESHOLD, int min_inf_bound = 0)#

Analysis of the format of a DiscreteParametric object.

Parameters:

error – [in] reference on a StatError object,
in_file – [in] stream,
line – [in] reference on the file line index,
last_ident – [in] identifier of the last distribution in the list,
cumul_threshold – [in] threshold on the cumulative distribution function,
min_inf_bound – [in] minimum lower bound.

Returns:

DiscreteParametric object.

static int nb_value_computation(discrete_parametric ident, int inf_bound, int sup_bound, double parameter, double probability, double cumul_threshold = CUMUL_THRESHOLD)#

Computation of the number of values of a parametric discrete distribution (binomial, Poisson, negative binomial, uniform, compound Poisson geometric, prior segment length distribution for a multiple change-point model).

Parameters:

ident – [in] distribution identifier,
inf_bound – [in] lower bound of the support,
sup_bound – [in] upper bound of the support (binomial or uniform distribution),
parameter – [in] parameter (negative binomial distribution),
probability – [in] probability (binomial or negative binomial distribution),
cumul_threshold – [in] threshold on the cumulative distribution function.

Returns:

number of values

class DiscreteParametricModel : public stat_tool::StatInterface, public stat_tool::DiscreteParametric #

Discrete parametric distribution with pointer on DiscreteDistributionData.

Public Functions

DiscreteParametricModel(const FrequencyDistribution &histo)#

Construction of a DiscreteParametricModel object from a FrequencyDistribution object.

Parameters:: histo – [in] reference on a FrequencyDistribution object.

DiscreteParametricModel(const Distribution &dist, const FrequencyDistribution *histo)#

Construction of a DiscreteParametricModel object from a Distribution object and a FrequencyDistribution object.

Parameters:

dist – [in] reference on a Distribution object,
histo – [in] pointer on a FrequencyDistribution object.

DiscreteParametricModel(const DiscreteParametric &dist, const FrequencyDistribution *histo)#

Construction of a DiscreteParametricModel object from a DiscreteParametric object and a FrequencyDistribution object.

Parameters:

dist – [in] reference on a DiscreteParametric object,
histo – [in] pointer on a FrequencyDistribution object.

DiscreteParametricModel(const DiscreteParametricModel &dist, bool data_flag = true)#

Constructor by copy of the DiscreteParametricModel class.

Parameters:

dist – [in] reference on a DiscreteParametricModel object,
data_flag – [in] flag copy of the DiscreteDistributionData object.

~DiscreteParametricModel()#: Destructor of the DiscreteParametricModel class.

DiscreteParametricModel &operator=(const DiscreteParametricModel &dist)#

Assignment operator of the DiscreteParametricModel class.

Parameters:: dist – [in] reference on a DiscreteParametricModel object.
Returns:: DiscreteParametricModel object.

DiscreteDistributionData *extract_data(StatError &error) const#

Extraction of the DiscreteDistributionData object included in a DiscreteParametricModel object.

Parameters:: error – [in] reference on a StatError object.
Returns:: DiscreteDistributionData object.

virtual std::ostream &line_write(std::ostream &os) const#

Writing on a single line of a DiscreteParametricModel object.

Parameters:: os – [inout] stream.

virtual bool plot_write(StatError &error, const char *prefix, const char *title = NULL) const#

Plot of a DiscreteParametricModel object using Gnuplot.

Parameters:

error – [in] reference on a StatError object,
prefix – [in] file prefix,
title – [in] figure title.

Returns:

error status.

virtual MultiPlotSet *get_plotable() const#

Plot of a DiscreteParametricModel object.

Returns:: MultiPlotSet object.

DiscreteDistributionData *simulation(StatError &error, int nb_element) const#

Building of a frequency distribution by simulating a discrete parametric distribution.

Parameters:

error – [in] reference on a StatError object,
nb_element – [in] sample size.

Returns:

DiscreteDistributionData object.

Public Static Functions

static DiscreteParametricModel *ascii_read(StatError &error, const std::string path, double cumul_threshold = CUMUL_THRESHOLD)#

Construction of a DiscreteParametricModel object from a file.

Parameters:

error – [in] reference on a StatError object,
path – [in] file path,
cumul_threshold – [in] threshold on the cumulative distribution function.

Returns:

DiscreteParametricModel object.

class DiscreteParametricProcess#

Discrete parametric observation process.

Collection of univariate discrete parametric emission distributions. In the multivariate case, use a collection of conditionally independent DiscreteParametricProcesses observations[k] is a pointer to emission distribution for state k

Public Functions

void copy(const DiscreteParametricProcess &process, bool mass_copy = false)#

Copy of a DiscreteParametricProcess object.

Parameters:

process – [in] reference on a DiscreteParametricProcess object.
mass_copy – [in] flag on copying or not probabilities (mass)

void remove()#: Destruction of the data members of a DiscreteParametricProcess object.

DiscreteParametricProcess(int inb_state = 0, int inb_value = 0)#

Constructor of the DiscreteParametricProcess class.

Parameters:

inb_state – [in] number of states,
inb_value – [in] number of values.

DiscreteParametricProcess(int inb_state, DiscreteParametric **pobservation)#

Constructor of the DiscreteParametricProcess class.

Parameters:

inb_state – [in] number of states,
pobservation – [in] pointer on the observation distributions.

~DiscreteParametricProcess()#: Destructor of the DiscreteParametricProcess class.

DiscreteParametricProcess &operator=(const DiscreteParametricProcess &process)#

Assignment operator of the DiscreteParametricProcess class.

Parameters:: process – [in] reference on a DiscreteParametricProcess object.
Returns:: DiscreteParametricProcess object.

Writing of a DiscreteParametricProcess object.

Parameters:

os – [inout] stream,
empirical_observation – [in] pointer on the observation frequency distributions,
marginal_distribution – [in] pointer on the marginal frequency distribution,
exhaustive – [in] flag detail level,
file_flag – [in] flag file,
model – [in] model type.

std::ostream &spreadsheet_print(std::ostream &os, FrequencyDistribution **empirical_observation = NULL, FrequencyDistribution *marginal_distribution = NULL, model_type model = HIDDEN_MARKOV) const#

Writing of a DiscreteParametricProcess object at the spreadsheet format.

Parameters:

os – [inout] stream,
empirical_observation – [in] pointer on the observation frequency distributions,
marginal_distribution – [in] pointer on the marginal frequency distribution,
model – [in] model type.

Plot of a DiscreteParametricProcess object using Gnuplot.

Parameters:

prefix – [in] file prefix,
title – [in] figure title,
process – [in] observation process index,
empirical_observation – [in] pointer on the observation frequency distributions,
marginal_distribution – [in] pointer on the marginal frequency distribution,
model – [in] model type.

Returns:

error status.

Plot of a DiscreteParametricProcess object.

Parameters:

plot – [in] reference on a MultiPlotSet object,
index – [in] MultiPlot index,
process – [in] observation process index,
empirical_observation – [in] pointer on the observation frequency distributions,
marginal_distribution – [in] pointer on the marginal frequency distribution,
model – [in] model type.

void nb_value_computation()#: Computation of the support of the observation process.

void state_permutation(int *permut) const#

Permutation of observation distributions. The permutation validity should be checked by the calling function.

Parameters:: permut – [in] permutation.

int nb_parameter_computation() const#

Computation of the number of free parameters of a discrete parametric observation process.

Returns:: number of free parameters.

double mean_computation(Distribution *pweight) const#

Computation of the mean of a mixture of discrete parametric observation distributions.

Parameters:: pweight – [in] pointer on the weight distribution.
Returns:: mixture mean.

double variance_computation(Distribution *pweight, double mean = D_INF) const#

Computation of the variance of a mixture of discrete parametric observation distributions.

Parameters:

pweight – [in] pointer on the weight distribution,
mean – [in] mean.

Returns:

mixture variance.

Distribution *mixture_computation(Distribution *pweight)#

Computation of a mixture of discrete parametric observation distributions.

Parameters:: pweight – [in] pointer on the weight distribution.
Returns:: mixture of discrete parametric observation distributions.

void init()#

Initialization of discrete parametric observation distributions using perturbation.

void DiscreteParametricProcess::init()

{ int i , j; double noise_proba , *pmass;

for (i = 0;i < nb_state;i++) { noise_proba = NOISE_PROBABILITY * nb_state / nb_value;

pmass = observation[i]->mass; for (j = 0;j < i * nb_value / nb_state;j++) { pmass++ -= noise_proba / (nb_state - 1); } for (j = i * nb_value / nb_state;j < (i + 1) * nb_value / nb_state;j++) { pmass++ += noise_proba; } for (j = (i + 1) * nb_value / nb_state;j < nb_value;j++) { pmass++ -= noise_proba / (nb_state - 1); } } }

Public Members

int nb_state#: number of states

int nb_value#: number of values

DiscreteParametric **observation#: discrete parametric observation distributions

Distribution *weight#: theoretical weights of observation distributions

Distribution *mixture#: mixture of observation distributions

Distribution *restoration_weight#: weights of observation distributions deduced from the restoration

Distribution *restoration_mixture#: mixture of observation distributions

Public Static Functions

static DiscreteParametricProcess *parsing(StatError &error, ifstream &in_file, int &line, int nb_state, model_type model, double cumul_threshold = OBSERVATION_THRESHOLD)#

Analysis of the format of discrete parametric observation distributions.

Parameters:

error – [in] reference on a StatError object,
in_file – [in] stream,
line – [in] reference on the file line index,
nb_state – [in] number of states,
model – [in] model type,
cumul_threshold – [in] threshold on the cumulative distribution functions.

Returns:

DiscreteParametricProcess object.

class DistanceMatrix : public stat_tool::StatInterface #

Distance matrix.

Subclassed by stat_tool::Clusters

Public Functions

DistanceMatrix()#: Default constructor of the DistanceMatrix class.

DistanceMatrix(int nb_pattern, const char *ilabel, int *pattern_identifier = NULL)#

Constructor of the DistanceMatrix class.

Parameters:

nb_pattern – [in] number of individuals,
ilabel – [in] label,
pattern_identifier – [in] individual identifiers,

DistanceMatrix(int nb_pattern, int irow_identifier, int icolumn_identifier, const char *ilabel, int *pattern_identifier = NULL, bool substitution_flag = true, bool transposition_flag = false)#

Constructor of the DistanceMatrix class.

Parameters:

nb_pattern – [in] number of individuals,
irow_identifier – [in] row identifier,
icolumn_identifier – [in] column identifier,
ilabel – [in] label,
pattern_identifier – [in] individual identifiers,
substitution_flag – [in] flag substitution edit operation,
transposition_flag – [in] flag transposition edit operation.

DistanceMatrix(const DistanceMatrix &dist_matrix, int inb_pattern, int *iidentifier, bool keep = true)#

Constructor of the DistanceMatrix class.

Parameters:

dist_matrix – [in] reference on a DistanceMatrix object,
inb_pattern – [in] number of selected individuals,
iidentifier – [in] identifiers of selected individuals,
keep – [in] flag for keeping or rejecting the selected individuals.

DistanceMatrix(const DistanceMatrix &dist_matrix, int nb_cluster, const char *ilabel)#

Constructor of the DistanceMatrix class.

Parameters:

dist_matrix – [in] reference on a DistanceMatrix object,
nb_cluster – [in] number of clusters,
ilabel – [in] label.

~DistanceMatrix()#: Destructor of the DistanceMatrix class.

DistanceMatrix &operator=(const DistanceMatrix &dist_matrix)#

Assignment operator of the DistanceMatrix class.

Parameters:: dist_matrix – [in] reference on a DistanceMatrix object.
Returns:: DistanceMatrix object.

DistanceMatrix *select_individual(StatError &error, int inb_pattern, int *iidentifier, bool keep = true) const#

Selection of individuals by their identifiers.

Parameters:

error – [in] reference on a StatError object,
inb_pattern – [in] number of individuals,
iidentifier – [in] individual identifiers,
keep – [in] flag for keeping or rejecting the selected individuals.

Returns:

DistanceMatrix object.

DistanceMatrix *symmetrize(StatError &error) const#

Symmetrization of a distance matrix.

Parameters:: error – [in] reference on a StatError object.
Returns:: DistanceMatrix object.

DistanceMatrix *unnormalize(StatError &error) const#

Unnormalization of a distance matrix.

Parameters:: error – [in] reference on a StatError object.
Returns:: DistanceMatrix object.

virtual std::ostream &line_write(std::ostream &os) const#

Writing on a single line of a DistanceMatrix object.

Parameters:: os – [inout] stream.

virtual bool plot_write(StatError &error, const char *prefix, const char *title = NULL) const#

Plot of a DistanceMatrix object using Gnuplot.

Parameters:

error – [in] reference on a StatError object,
prefix – [in] file prefix,
title – [in] figure title.

Returns:

error status.

virtual MultiPlotSet *get_plotable() const#

Plot of a DistanceMatrix object.

Returns:: MultiPlotSet object.

bool test_symmetry() const#

Test of the symmetry of a distance matrix.

Returns:: flag symmetric matrix or not.

void update(int irow_identifier, int icolumn_identifier, double idistance, int alignment_length, double ideletion_distance, int inb_deletion, double iinsertion_distance, int inb_insertion, int inb_match, double isubstitution_distance = 0., int inb_substitution = 0, double itransposition_distance = 0., int inb_transposition = 0)#

Update of a DistanceMatrix object with the result of the alignment of 2 structures.

Parameters:

irow_identifier – [in] row identifier (1st individual),
icolumn_identifier – [in] column identifier (2nd individual),
idistance – [in] distance,
alignment_length – [in] alignment length,
ideletion_distance – [in] deletion distance,
inb_deletion – [in] number of deletions,
iinsertion_distance – [in] insertion distance,
inb_insertion – [in] number of insertions,
inb_match – [in] number of matches,
isubstitution_distance – [in] substitution distance,
inb_substitution – [in] number of substitutions,
itransposition_distance – [in] transposition distance,
inb_transposition – [in] number of transpositions.

void update(int irow_identifier, int icolumn_identifier, double idistance, int ilength)#

Update of a DistanceMatrix object with the result of a comparison.

Parameters:

irow_identifier – [in] row identifier (1st individual),
icolumn_identifier – [in] column identifier (2nd individual),
idistance – [in] distance,
ilength – [in] length.

Dendrogram *agglomerative_hierarchical_clustering(hierarchical_strategy strategy, linkage criterion = AVERAGE_NEIGHBOR) const#

Agglomerative hierarchical clustering algorithm (Kaufman & Rousseeuw, pp. 199-208).

Parameters:

strategy – [in] algorithm type,
criterion – [in] cluster merging criterion.

Returns:

Dendrogram object.

Dendrogram *divisive_hierarchical_clustering() const#

Divisive hierarchical clustering algorithm (Kaufman & Rousseeuw, pp. 253-259).

Returns:: Dendrogram object.

bool hierarchical_clustering(StatError &error, std::ostream *os, hierarchical_strategy strategy = AGGLOMERATIVE, linkage criterion = AVERAGE_NEIGHBOR, const std::string path = "", output_format format = ASCII) const#

Hierarchical clustering algorithms.

Parameters:

error – [in] reference on a StatError object,
os – [in] stream for displaying the hierarchical clustering results,
strategy – [in] algorithm type (AGGLOMERATIVE/DIVISIVE/ORDERING),
criterion – [in] cluster merging criterion (agglomerative algorithm),
path – [in] file path,
format – [in] output format (ASCII/SPREADSHEET).

Returns:

Dendrogram object.

class Distribution#

Discrete distribution.

Probabilities are represented as an array *mass, with size nb_allocated_value mass[i] is only meaningful for offset <= i < nb_value. For i < offset or i >= nb_value, mass[i] may be either 0 or unspecified. nb_allocated_value has to be greater or equal to nb_value. It may happen that nb_allocated_value > nb_value, particularly when mass[i] is computed for every i <= quantile(0.9999) = nb_value-1, which we do not know in advance, but we allocated nb_allocated_value > nb_value for the sake of safety.

Subclassed by stat_tool::Compound, stat_tool::Convolution, stat_tool::DiscreteMixture, stat_tool::DiscreteParametric

Public Functions

void mass_copy(const Distribution &dist, int inb_value = I_DEFAULT)#

Copy of the probability mass function.

Parameters:

dist – [in] reference on a Distribution object,
inb_value – [in] number of values. inb_value can be less than dist.nb_values. In this case, the unnormalized truncated distribution is copied

void equal_size_copy(const Distribution &dist)#

Copy of a Distribution object in the case where the number of allocated values is the same for the 2 Distribution objects.

Parameters:: dist – [in] reference on a Distribution object.

void init(int inb_value)#

Initialization of a Distribution object.

Parameters:: inb_value – [in] number of values.

void copy(const Distribution &dist, int ialloc_nb_value = I_DEFAULT)#

Copy of a Distribution object.

Parameters:

dist – [in] reference on a Distribution object,
ialloc_nb_value – [in] number of allocated values.

void normalization_copy(const Distribution &dist)#

Copy of a Distribution object with renormalization.

Parameters:: dist – [in] reference on a Distribution object.

Distribution(int inb_value = 0)#

Constructor of the Distribution class.

Parameters:: inb_value – [in] number of values.

Distribution(int inb_value, double *imass)#

Constructor of the Distribution class.

Parameters:

inb_value – [in] number of values,
imass – [in] probability mass function.

Distribution(const Distribution &dist, double scaling_coeff)#

Construction of a Distribution object from an initial Distribution object applying a scaling operation.

Parameters:

dist – [in] reference on a Distribution object,
scaling_coeff – [in] scaling factor.

Distribution(const FrequencyDistribution &histo)#

Construction of a Distribution object from a FrequencyDistribution object.

Parameters:: histo – [in] reference on a FrequencyDistribution object.

Distribution(const Distribution &dist, distribution_transformation transform = DISTRIBUTION_COPY, int ialloc_nb_value = I_DEFAULT)#

Copy constructor of the Distribution class.

Parameters:

dist – [in] reference on a Distribution object,
transform – [in] type of transform (DISTRIBUTION_COPY/NORMALIZATION),
ialloc_nb_value – [in] number of allocated values.

virtual ~Distribution()#: Destructor of the Distribution class.

Distribution &operator=(const Distribution&)#

Assignment operator of the Distribution class.

Parameters:: dist – [in] reference on a Distribution object.
Returns:: Distribution object.

bool operator==(const Distribution&) const#

Equality operator of the Distribution class.

Parameters:: dist – [in] reference on a Distribution object.
Returns:: equality or not of the discrete distributions.

std::ostream &ascii_characteristic_print(std::ostream &os, bool shape = false, bool comment_flag = false) const#

Writing of the characteristics of a discrete distribution.

Parameters:

os – [inout] stream,
shape – [in] flag on the writing of the shape characteristics,
comment_flag – [in] flag comment.

std::ostream &print(std::ostream &os) const#

Display of a discrete distribution.

Parameters:: os – [inout] stream.

std::ostream &spreadsheet_characteristic_print(std::ostream &os, bool shape = false) const#

Writing of the characteristics of a discrete distribution at the spreadsheet format.

Parameters:

os – [inout] stream,
shape – [in] flag on the writing of the shape characteristics.

int plot_nb_value_computation(const FrequencyDistribution *histo = NULL) const#

Computation of the number of plotted values (Gnuplot output).

Parameters:: histo – [in] pointer a frequency distribution.
Returns:: number of plotted values.

bool plot_print(const char *path, double *concentration, double scale) const#

Writing of a discrete distribution at the Gnuplot format.

Parameters:

path – [in] file path,
concentration – [in] pointer on the concentration function,
scale – [in] scaling factor.

Returns:

error status.

bool plot_print(const char *path, const FrequencyDistribution *histo = NULL) const#

Writing of a discrete distribution and a frequency distribution at the Gnuplot format.

Parameters:

path – [in] file path,
histo – [in] pointer on a frequency distribution.

Returns:

error status.

bool survival_plot_print(const char *path, double *survivor) const#

Writing of a discrete distribution and the associated survivor function at the Gnuplot format.

Parameters:

path – [in] file path,
survivor – [in] pointer on the survivor function.

Returns:

error status.

bool plot_write(StatError &error, const char *prefix, int nb_dist, const Distribution **idist, const char *title) const#

Plot of family of a discrete distributions using Gnuplot:

probability mass functions and cumulative distribution functions,
matching of cumulative distribution functions,
concentration curves.

Parameters:

error – [in] reference on a StatError object,
prefix – [in] file prefix,
nb_dist – [in] number of distributions,
idist – [in] pointer on the discrete distributions,
title – [in] figure title.

Returns:

error status.

void plotable_mass_write(SinglePlot &plot, double scale = 1.) const#

Writing of the probability mass function.

Parameters:

plot – [in] reference on a SinglePlot object,
scale – [in] scaling factor (default value: 1).

void plotable_cumul_write(SinglePlot &plot) const#

Writing of the cumulative discrete distribution.

Parameters:: plot – [in] reference on a SinglePlot object.

void plotable_cumul_matching_write(SinglePlot &plot, const Distribution &reference_dist) const#

Writing of the matching of the cumulative distribution function with the cumulative distribution function of a reference distribution.

Parameters:

plot – [in] reference on a SinglePlot object
reference_dist – [in] reference on the reference distribution.

void plotable_concentration_write(SinglePlot &plot) const#

Writing of the concentration curve deduced from a discrete distribution.

Parameters:: plot – [in] reference on a SinglePlot object.

void plotable_survivor_write(SinglePlot &plot) const#

Computation and writing of the survivor function of a discrete distribution.

Parameters:: plot – [in] reference on a SinglePlot object.

MultiPlotSet *get_plotable() const#

Plot of a discrete distribution.

Returns:: MultiPlotSet object.

MultiPlotSet *get_plotable_distributions(StatError &error, int nb_dist, const Distribution **idist) const#

Plot of family of a discrete distributions:

probability mass functions and cumulative distribution functions,
matching of cumulative distribution functions,
concentration curves.

Parameters:

error – [in] reference on a StatError object,
nb_dist – [in] number of distributions,
idist – [in] pointer on the discrete distributions.

Returns:

MultiPlotSet object.

std::string survival_ascii_write() const#

Computation of the survival rates from a discrete distribution and writing of the result.

Returns:: string.

bool survival_spreadsheet_write(StatError &error, const std::string path) const#

Computation of survival rates from a discrete distribution and writing of the result in a file at the spreadsheet format.

Parameters:

error – [in] reference on a StatError object,
path – [in] file path.

Returns:

error status.

bool survival_plot_write(StatError &error, const char *prefix, const char *title = NULL) const#

Computation of the survival rates from a discrete distribution and plot of the result using Gnuplot.

Parameters:

error – [in] reference on a StatError object,
prefix – [in] file prefix,
title – [in] figure title.

Returns:

error status.

MultiPlotSet *survival_get_plotable(StatError &error) const#

Computation of survival rates from a discrete distribution and plot of the result.

Parameters:: error – [in] reference on a StatError object.
Returns:: MultiPlotSet object.

void max_computation()#: Extraction of the probability at the mode of a discrete distribution.

double mode_computation() const#

Extraction of the mode of a discrete distribution.

Returns:: mode.

void mean_computation()#: Computation of the mean of a discrete distribution.

double quantile_computation(double icumul = 0.5) const#

Computation of a quantile of a discrete distribution.

Parameters:: icumul – [in] value of the cumulative distribution function.
Returns:: quantile.

void variance_computation()#: Computation of the variance of a discrete distribution.

void nb_value_computation()#: Computation of the number of possible values from 0.

void offset_computation()#: Computation of the number of values of null probability from 0.

double concentration_computation() const#

Computation of the coefficient of concentration of a discrete distribution.

Returns:: coefficient of concentration.

double mean_absolute_deviation_computation(double location) const#

Computation of the mean absolute deviation of a discrete distribution.

Parameters:: location – [in] location measure (e.g. mean or median).
Returns:: mean absolute deviation.

double skewness_computation() const#

Computation of the coefficient of skewness of a discrete distribution.

Returns:: coefficient of skewness.

double kurtosis_computation() const#

Computation of the excess kurtosis of a discrete distribution: excess kurtosis = coefficient of kurtosis - 3.

Returns:: excess kurtosis.

double information_computation() const#

Computation of the information quantity of a discrete distribution.

Returns:: information quantity.

double first_difference_norm_computation() const#

Computation of the sum of squared first-order differences.

Returns:: sum of squared first-order differences.

double second_difference_norm_computation() const#

Computation of the sum of squared second-order differences.

Returns:: sum of squared second-order differences.

void cumul_computation()#: Computation of the cumulative distribution function of a discrete distribution.

double *survivor_function_computation() const#

Computation of the survivor function of a discrete distribution.

Returns:: survivor function.

double *concentration_function_computation() const#

Computation of the concentration function of a discrete distribution.

Returns:: concentration function.

double overlap_distance_computation(const Distribution &dist) const#

Computation of the distance between 2 discrete distributions (1 - overlap).

Returns:: overlap distance.

void log_computation()#: Ccomputation of the log-probability mass function.

double survivor_likelihood_computation(const FrequencyDistribution &histo) const#

Computation of the log-likelihood of the survival function of a given distribution.

Parameters:: histo – [in] reference on a FrequencyDistribution object.
Returns:: log-likelihood.

double chi2_value_computation(const FrequencyDistribution &histo) const#

Computation of the chi2 value for a discrete distribution fit (Chi2 goodness of fit test).

Parameters:: histo – [in] reference on a FrequencyDistribution object.
Returns:: chi2 value.

void chi2_degree_of_freedom(const FrequencyDistribution &histo, Test &test) const#

Grouping of successive values, computation of the degrees of freedom and the chi2 value (Chi2 goodness of fit test).

Parameters:

histo – [in] reference on a FrequencyDistribution object,
test – [in] reference on a Test object.

void penalty_computation(double weight, penalty_type pen_type, double *penalty, side_effect outside) const#

Computation of the penalty terms in the framework of a penalized likelihood approach.

Parameters:

weight – [in] penalty weight,
pen_type – [in] penalty type (1st order, 2nd order difference or entropy),
penalty – [in] penalties,
outside – [in] management of side effects (zero outside the support or continuation of the distribution).

void chi2_fit(const FrequencyDistribution &histo, Test &test) const#

Chi2 goodness of fit test for a discrete distribution.

Parameters:

histo – [in] reference on a FrequencyDistribution object,
test – [in] reference on a Test object.

void convolution(Distribution &dist1, Distribution &dist2, int inb_value = I_DEFAULT)#

Convolution of 2 distributions (the convolution can be put in one of the 2 distributions).

Parameters:

dist1 – [in] reference on the 1st distribution,
dist2 – [in] reference on the 2nd distribution,
inb_value – [in] number of values of the convolution.

int simulation() const#

Simulation using the cumulative distribution function.

Returns:: generated value.

DiscreteParametricModel *truncate(StatError &error, int imax_value) const#

Truncation of a distribution.

Parameters:

error – [in] reference on a StatError object,
imax_value – [in] maximum value.

Returns:

DiscreteParametricModel object.

Public Members

int nb_value#: number of values from 0

int alloc_nb_value#: number of allocated values

int offset#: number of values of null probability from 0

double max#: maximum probability

double complement#: complementary probability (> 0. for improper distributions)

double mean#: mean

double variance#: variance

int nb_parameter#: number of free parameters

double *mass#: probability mass function

double *cumul#: cumulative distribution function

class Forward : public stat_tool::DiscreteParametric #

Forward recurrence or sojourn time distribution.

Public Functions

void computation(const DiscreteParametric &dist)#

Computation of the forward recurrence or sojourn time distribution on the basis of the recurrence or sojourn time distribution.

Parameters:: dist – [in] recurrence or sojourn time distribution.

class FrequencyDistribution : public stat_tool::Reestimation<int>#

Frequency distribution.

Subclassed by stat_tool::CompoundData, stat_tool::ConvolutionData, stat_tool::DiscreteDistributionData, stat_tool::DiscreteMixtureData

Public Functions

FrequencyDistribution(int inb_element, int *ielement)#

Constructor of the FrequencyDistribution class.

Parameters:

inb_element – [in] number of individuals,
ielement – [in] individuals.

FrequencyDistribution(const FrequencyDistribution &histo, frequency_distribution_transformation transform, int param, rounding mode = FLOOR)#

Constructor of the FrequencyDistribution class.

Parameters:

histo – [in] reference on a FrequencyDistribution object,
transform – [in] type of transform (SHIFT/CLUSTER),
param – [in] shifting parameter (SHIFT) / clustering step (CLUSTER),
mode – [in] clustering mode (FLOOR/ROUND/CEIL).

bool operator==(const FrequencyDistribution&) const#

Equality operator of the FrequencyDistribution class.

Parameters:: histo – [in] reference on a FrequencyDistribution object.
Returns:: equality or not of the frequency distributions.

std::ostream &ascii_print(std::ostream &os, int comment_flag = false, bool cumul_flag = false) const#

Writing of a frequency distribution.

Parameters:

os – [inout] stream,
comment_flag – [in] flag comment,
cumul_flag – [in] flag on the writing of the cumulative distribution function.

std::ostream &spreadsheet_characteristic_print(std::ostream &os, bool shape = false) const#

Writing of the characteristics of a frequency distribution at the spreadsheet format.

Parameters:

os – [inout] stream,
shape – [in] flag on the writing of the shape characteristics.

std::ostream &spreadsheet_circular_characteristic_print(std::ostream &os) const#

Writing of the characteristics of a frequency distribution of a circular variable at the spreadsheet format.

Parameters:: os – [inout] stream.

std::ostream &spreadsheet_print(std::ostream &os, bool cumul_flag = false, bool concentration_flag = false) const#

Writing of a frequency distribution at the spreadsheet format.

Parameters:

os – [inout] stream,
cumul_flag – [in] flag on the writing of the cumulative distribution function,
concentration_flag – [in] flag on the writing of the concentration function.

bool plot_print(const char *path, int nb_histo = 0, const FrequencyDistribution **histo = NULL) const#

Writing of a family of frequency distributions at the Gnuplot format.

Parameters:

path – [in] file path,
nb_histo – [in] number of frequency distributions,
histo – [in] pointer on the frequency distributions.

Returns:

error status.

bool plot_print(const char *path, double *cumul, double *concentration, double shift = 0.) const#

Writing of a frequency distribution at the Gnuplot format.

Parameters:

path – [in] file path,
cumul – [in] pointer on the cumulative distribution function,
concentration – [in] pointer on the concentration function,
shift – [in] shift of the frequency distribution.

Returns:

error status.

bool survival_plot_print(const char *path, double *survivor) const#

Writing of a frequency distribution, the deduced probability mass and survivor functions at the Gnuplot format.

Parameters:

path – [in] file path,
survivor – [in] pointer on the survivor function.

Returns:

error status.

void plotable_frequency_write(SinglePlot &plot) const#

Writing of a frequency distribution.

Parameters:: plot – [in] reference on a SinglePlot object.

void plotable_mass_write(SinglePlot &plot) const#

Writing of the probability mass function deduced from a frequency distribution.

Parameters:: plot – [in] reference on a SinglePlot object.

void plotable_cumul_write(SinglePlot &plot, double *icumul = NULL, double scale = D_DEFAULT) const#

Writing of the cumulative distribution function deduced from a frequency distribution.

Parameters:

plot – [in] reference on a SinglePlot object,
icumul – [in] pointer on the cumulative distribution function,
scale – [in] scaling factor.

void plotable_cumul_matching_write(SinglePlot &plot, int reference_offset, int reference_nb_value, double *reference_cumul, double *icumul = NULL) const#

Writing of the matching of a cumulative distribution function with a reference cumulative distribution function.

Parameters:

plot – [in] reference on a SinglePlot object,
reference_offset – [in] reference minimum value,
reference_nb_value – [in] reference number of values,
reference_cumul – [in] pointer on the reference cumulative distribution function,
icumul – [in] pointer on the cumulative distribution function.

void plotable_concentration_write(SinglePlot &plot, double *icumul = NULL, double scale = D_DEFAULT) const#

Writing of the concentration curve deduced from a frequency distribution.

Parameters:

plot – [in] reference on a SinglePlot object,
icumul – [in] pointer on the cumulative distribution function,
scale – [in] scaling factor.

void plotable_survivor_write(SinglePlot &plot) const#

Computation and writing of the survivor function computed from a frequency distribution.

Parameters:: plot – [in] reference on a SinglePlot object.

double *cumul_computation(double scale = D_DEFAULT) const#

Computation de la cumulative distribution function deduced from a frequency distribution.

Parameters:: scale – [in] scaling factor.
Returns:: cumulative distribution function.

double concentration_computation() const#

Computation of the coefficient of concentration of a frequency distribution.

Returns:: coefficient of concentration.

double *survivor_function_computation(double scale = D_DEFAULT) const#

Computation of the survivor function deduced from a frequency distribution.

Parameters:: scale – [in] scaling factor.
Returns:: survivor function.

double *concentration_function_computation(double scale = D_DEFAULT) const#

Computation of the concentration function deduced from a frequency distribution.

Parameters:: scale – [in] scaling factor.
Returns:: concentration function.

Test *kruskal_wallis_test(int nb_histo, const FrequencyDistribution **ihisto) const#

Kruskal-Wallis test (analysis of variance on ranks).

Parameters:

nb_histo – [in] number of frequency distributions,
ihisto – [in] pointer on the frequency distributions.

Returns:

Test object.

bool dissimilarity_spreadsheet_write(StatError &error, const std::string path, int nb_histo, const FrequencyDistribution **ihisto, variable_type type, double **dissimilarity) const#

Writing of the results of a comparison of frequency distributions in a file at the spreadsheet format.

Parameters:

error – [in] reference on a StatError object,
path – [in] file path,
nb_histo – [in] number of frequency distributions,
ihisto – [in] pointer on the frequency distributions,
type – [in] variable type (NOMINAL/ORDINAL/NUMERIC),
dissimilarity – [in] dissimilarities.

Returns:

error status.

void update(const Reestimation<double> *reestim, int inb_element)#

Update of an integer frequency distribution from a real frequency distribution by rounding.

Parameters:

reestim – [in] pointer on the real frequency distribution,
inb_element – [in] cumulative frequencies.

FrequencyDistribution *frequency_scale(int inb_element) const#

Computation of a frequency distribution from an initial frequency distribution by scaling frequencies.

Parameters:: inb_element – [in] cumulative frequencies.
Returns:: scaled frequency distribution.

double *rank_computation() const#

Computation of ranks of values on the basis of a frequency distribution.

Returns:: ranks.

int cumulative_distribution_function_computation(double **cdf) const#

Computation of the cumulative frequency distribution function.

Parameters:: cdf – [in] (values, cumulative frequency distribution function).
Returns:: number of values.

int min_interval_computation() const#

Computation of the minimum interval between 2 values.

Returns:: minimum interval.

DiscreteParametric *parametric_estimation(discrete_parametric ident, int min_inf_bound = 0, bool flag = true, double cumul_threshold = CUMUL_THRESHOLD) const#

Estimation of a discrete parametric distribution (binomial, Poisson or negative binomial).

Parameters:

ident – [in] distribution identifier,
min_inf_bound – [in] minimum lower bound,
flag – [in] flag on the estimation of the lower bound,
cumul_threshold – [in] threshold on the cumulative distribution function.

Returns:

DiscreteParametric object.

double likelihood_computation(const ContinuousParametric &dist, int min_interval = I_DEFAULT) const#

Computation of the log-likelihood of a continuous distribution for a frequency distribution.

Parameters:

dist – [in] reference on a ContinuousParametric object,
min_interval – [in] minimum interval between 2 values.

Returns:

log-likelihood.

DiscreteDistributionData *merge(int nb_sample, const std::vector<FrequencyDistribution> &ihisto) const#

Merging of FrequencyDistribution objects.

Parameters:

nb_sample – [in] number of FrequencyDistribution objects,
ihisto – [in] pointer on the FrequencyDistribution objects.

Returns:

DiscreteDistributionData object.

void shift(const FrequencyDistribution &histo, int shift_param)#

Shifting of a frequency distribution.

Parameters:

histo – [in] reference on a FrequencyDistribution object,
shift_param – [in] shifting parameter.

void cluster(const FrequencyDistribution &histo, int step, rounding mode)#

Clustering of values of a frequency distribution.

Parameters:

histo – [in] reference on a FrequencyDistribution object,
step – [in] clustering step,
mode – [in] mode (FLOOR/ROUND/CEIL).

DiscreteDistributionData *shift(StatError &error, int shift_param) const#

Shifting of a frequency distribution.

Parameters:

error – [in] reference on a StatError object,
shift_param – [in] shifting parameter.

Returns:

DiscreteDistributionData object.

DiscreteDistributionData *cluster(StatError &error, int step, rounding mode = FLOOR) const#

clustering of values of a frequency distribution.

Parameters:

error – [in] reference on a StatError object,
step – [in] clustering step,
mode – [in] mode (FLOOR/ROUND/CEIL).

Returns:

DiscreteDistributionData object.

DiscreteDistributionData *cluster(StatError &error, int nb_class, int *ilimit) const#

Partitioning of values of a frequency distribution.

Parameters:

error – [in] reference on a StatError object,
nb_class – [in] number of classes,
ilimit – [in] limits between classes (beginning of classes).

Returns:

DiscreteDistributionData object.

DiscreteDistributionData *transcode(StatError &error, int *category) const#

Transcoding of categories.

Parameters:

error – [in] reference on a StatError object,
category – [in] transcoding table.

Returns:

DiscreteDistributionData object.

DiscreteDistributionData *value_select(StatError &error, int min_value, int max_value, bool keep = true) const#

Selection of a range of values.

Parameters:

error – [in] reference on a StatError object,
min_value – [in] lowest value,
max_value – [in] highest value,
keep – [in] flag for keeping or rejecting the selected values.

Returns:

DiscreteDistributionData object.

bool plot_write(StatError &error, const char *prefix, int nb_histo, const FrequencyDistribution **ihisto, const char *title) const#

Plot of a family of frequency distributions using Gnuplot:

frequency distributions and cumulative frequencies,
probability mass functions and cumulative distribution functions,
matching of cumulative distribution functions,
concentration curves.

Parameters:

error – [in] reference on a StatError object,
prefix – [in] file prefix,
nb_histo – [in] number of frequency distributions,
ihisto – [in] pointer on the frequency distributions,
title – [in] figure title.

Returns:

error status.

MultiPlotSet *get_plotable() const#

Plot of a frequency distribution.

Returns:: MultiPlotSet object.

MultiPlotSet *get_plotable_frequency_distributions(StatError &error, int nb_histo, const FrequencyDistribution **ihisto) const#

Plot of a family of frequency distributions:

frequency distributions and cumulative frequencies,
probability mass functions and cumulative distribution functions,
matching of cumulative distribution functions,
concentration curves.

Parameters:

error – [in] reference on a StatError object,
nb_histo – [in] number of frequency distributions,
ihisto – [in] pointer on the frequency distributions.

Returns:

plots.

std::string survival_ascii_write() const#

Computation of the survival rates from a frequency distribution and writing of the result.

Returns:: string.

bool survival_spreadsheet_write(StatError &error, const std::string path) const#

Computation of the survival rates from a frequency distribution and writing of the result in a file at the spreadsheet format.

Parameters:

error – [in] reference on a StatError object,
path – [in] file path.

Returns:

error status.

bool survival_plot_write(StatError &error, const char *prefix, const char *title = NULL) const#

Computation of the survival rates from a frequency distribution and plot of the result using Gnuplot.

Parameters:

error – [in] reference on a StatError object,
prefix – [in] file prefix,
title – [in] figure title.

Returns:

error status.

MultiPlotSet *survival_get_plotable(StatError &error) const#

Computation of the survival rates from a frequency distribution and plot of the result.

Parameters:: error – [in] reference on a StatError object.
Returns:: plots.

void F_comparison(std::ostream &os, const FrequencyDistribution &histo) const#

F test of variance comparison.

Parameters:

os – [in] stream for displaying the test results,
histo – [in] reference on a frequency distribution.

void t_comparison(std::ostream &os, const FrequencyDistribution &histo) const#

Student’s t test of mean comparison.

Parameters:

os – [in] stream for displaying the test results,
histo – [in] reference on a frequency distribution.

bool wilcoxon_mann_whitney_comparison(StatError &error, std::ostream &os, const FrequencyDistribution &ihisto) const#

Wilcoxon-Mann-Whitney test of distribution comparison.

Parameters:

error – [in] reference on a StatError object,
os – [in] stream for displaying the test results,
ihisto – [in] reference on a frequency distribution.

Returns:

error status.

DiscreteParametricModel *fit(StatError &error, const DiscreteParametric &idist) const#

Fit of a distribution.

Parameters:

error – [in] reference on a StatError object,
idist – [in] reference on a DiscreteParametric object.

Returns:

DiscreteParametricModel object.

DiscreteParametricModel *parametric_estimation(StatError &error, discrete_parametric ident, int min_inf_bound = 0, bool flag = true, double cumul_threshold = CUMUL_THRESHOLD) const#

Estimation of a discrete parametric distribution (binomial, Poisson or negative binomial).

Parameters:

error – [in] reference on a StatError object,
ident – [in] distribution identifier,
min_inf_bound – [in] minimum lower bound,
flag – [in] flag on the estimation of the lower bound,
cumul_threshold – [in] threshold on the cumulative distribution function.

Returns:

DiscreteParametricModel object.

DiscreteParametricModel *type_parametric_estimation(StatError &error, int min_inf_bound = 0, bool flag = true, double cumul_threshold = CUMUL_THRESHOLD) const#

Estimation of a discrete parametric distribution (binomial, Poisson or negative binomial).

Parameters:

error – [in] reference on a StatError object,
min_inf_bound – [in] minimum lower bound,
flag – [in] flag on the estimation of the lower bound,
cumul_threshold – [in] threshold on the cumulative distribution function.

Returns:

DiscreteParametricModel object.

DiscreteMixture *discrete_mixture_estimation(StatError &error, const DiscreteMixture &imixt, bool *estimate, int min_inf_bound = 0, bool mixt_flag = true, bool component_flag = true, double weight_step = 0.1) const#

Estimation of a mixture of parametric discrete distributions using the EM algorithm.

Parameters:

error – [in] reference on a StatError object,
imixt – [in] pointer on the known components,
estimate – [in] flags on the known components,
min_inf_bound – [in] minimum lower bound of the mixture support,
mixt_flag – [in] flag on the lower bound of the mixture support,
component_flag – [in] flag on the lower bounds of the component supports,
weight_step – [in] step for weight initialization.

Returns:

DiscreteMixture object.

DiscreteMixture *discrete_mixture_estimation(StatError &error, const DiscreteMixture &imixt, int min_inf_bound = 0, bool mixt_flag = true, bool component_flag = true, double weight_step = 0.1) const#

Estimation of a mixture of parametric discrete distributions using the EM algorithm.

Parameters:

error – [in] reference on a StatError object,
imixt – [in] pointer on the known components,
min_inf_bound – [in] minimum lower bound of the mixture support,
mixt_flag – [in] flag on the lower bound of the mixture support,
component_flag – [in] flag on the lower bounds of the component supports,
weight_step – [in] step for weight initialization.

Returns:

DiscreteMixture object.

DiscreteMixture *discrete_mixture_estimation(StatError &error, int nb_component, discrete_parametric *ident, int min_inf_bound = 0, bool mixt_flag = true, bool component_flag = true, double weight_step = 0.1) const#

Estimation of a mixture of parametric discrete distributions using the EM algorithm.

Parameters:

error – [in] reference on a StatError object,
nb_component – [in] number of components,
ident – [in] component identifiers,
min_inf_bound – [in] minimum lower bound of the mixture support,
mixt_flag – [in] flag on the lower bound of the mixture support,
component_flag – [in] flag on the lower bounds of the component supports,
weight_step – [in] step for weight initialization.

Returns:

DiscreteMixture object.

class Histogram#

Histogram.

Histogram are represented as an array *frequency, with size nb_bin

Public Functions

void copy(const Histogram &histo)#

Copy of an Histogram object.

Parameters:: histo – [in] reference on an Histogram object.

Histogram(int inb_bin = 0, bool init_flag = true)#

Construction of an Histogram object.

Parameters:

inb_bin – [in] number of bins,
init_flag – [in] flag initialization.

Histogram(const FrequencyDistribution &histo)#

Construction of an Histogram object from a FrequencyDistribution object.

Parameters:: histo – [in] reference on a FrequencyDistribution object.

~Histogram()#: Destructor of the Histogram class.

Histogram &operator=(const Histogram &histo)#

Assignment operator of the Histogram class.

Parameters:: histo – [in] reference on an Histogram object.
Returns:: Histogram object.

std::ostream &ascii_print(std::ostream &os, bool comment_flag = false) const#

Writing of an histogram.

Parameters:

os – [inout] stream,
comment_flag – [in] flag comment.

std::ostream &spreadsheet_print(std::ostream &os) const#

Writing of an histogram at the spreadsheet format.

Parameters:: os – [inout] stream.

bool plot_print(const char *path) const#

Writing of an histogram at the Gnuplot format.

Parameters:: path – [in] file path.
Returns:: error status.

void plotable_write(SinglePlot &plot) const#

Writing of an histogram at the plotable format.

Parameters:: plot – [in] reference on a SinglePlot object.

void max_computation()#: Determination of the maximum bin frequency for an histogram.

double *cumul_computation() const#

Computation of a cumulative distribution function from an histogram.

Returns:: cumulative distribution function.

Public Members

int nb_element#: sample size

int nb_bin#: number of bins

double bin_width#: constant bin width

int max#: maximum value in frequency

int *frequency#: frequency for each bin

int type#: variable type (INT_VALUE/REAL_VALUE)

double min_value#: minimum value of support

double max_value#: maximum value of support

class MultiPlot#

class MultivariateMixture : public stat_tool::StatInterface #

Emission distributions are represented by arrays of DiscreteParametricProcess* and CategoricalProcess*. pcomponent[var] == NULL if and only if npcomponent[var] != NULL in which case the variable is represented by a categorical distribution for each mixture components

Public Functions

DiscreteParametricModel *extract_parametric_model(StatError &error, int ivariable, int index) const#: extract parametric component

Distribution *extract_categorical_model(StatError &error, int ivariable, int index) const#: extract categorical component

DiscreteMixture *extract_distribution(StatError &error, int ivariable) const#: extract marginal mixture distribution

void state_permutation(StatError &error, int *perm) const#: Permutation of the states of self

virtual MultiPlotSet *get_plotable() const#

MultivariateMixtureData *cluster(StatError &error, const Vectors &vec, int algorithm = VITERBI, bool add_state_entropy = false) const#: add restored states and state entropy to Vectors

bool is_parametric(int ivariable) const#: return “true” if process ivariable is parametric

class MultivariateMixtureData : public stat_tool::Vectors #

template<typename Type> class Reestimation#

Frequency distribution with integer or real (for EM algorithms) frequencies.

Frequencies are represented as an array *frequency, with size alloc_nb_value frequency[i] is only meaningful for offset <= i < nb_value. For i < offset, frequency[i] may be either 0 or unspecified.

Public Functions

void init(int inb_value)#

Construction of a Reestimation object.

Parameters:: inb_value – [in] number of values from 0.

void copy(const Reestimation<Type> &histo)#

Copy of a Reestimation object.

Parameters:: histo – [in] reference on a Reestimation object.

Reestimation(const Reestimation<Type> &histo)#

Constructor by copy of the Reestimation class.

Parameters:: histo – [in] reference on a Reestimation object.

Reestimation(int nb_histo, const Reestimation<Type> **histo)#

Constructor by merging of the Reestimation class.

Parameters:

nb_histo – [in] number of Reestimation objects,
histo – [in] pointer on the Reestimation objects.

~Reestimation()#: Destructor of the Reestimation class.

Reestimation<Type> &operator=(const Reestimation<Type> &histo)#

Assignment operator of the Reestimation class.

Parameters:: histo – [in] reference on a Reestimation object.
Returns:: Reestimation object.

std::ostream &ascii_characteristic_print(std::ostream &os, bool shape = false, bool comment_flag = false) const#

Writing of the characteristics of a Reestimation object.

Parameters:

os – [inout] stream,
shape – [in] flag on the writing of the shape characteristics,
comment_flag – [in] flag comments.

std::ostream &ascii_circular_characteristic_print(std::ostream &os, bool comment_flag = false) const#

Writing of the characteristics of a Reestimation object for a circular variable.

Parameters:

os – [inout] stream,
comment_flag – [in] flag comment.

std::ostream &print(std::ostream &os) const#

Display of a Reestimation object.

Parameters:: os – [inout] stream.

void nb_value_computation()#: Computation of the number of values from 0.

void offset_computation()#: Computation of the number of values of null frequency from 0.

void nb_element_computation()#: Computation of the sample size.

void max_computation()#: Determination of the maximum frequency.

double mode_computation() const#

Determination of the mode.

Returns:: mode.

void mean_computation()#: Mean computation.

double quantile_computation(double icumul = 0.5) const#

Computation of a quantile.

Parameters:: icumul – [in] value of the cumulative distribution function.
Returns:: quantile.

void variance_computation(bool bias = false)#

Variance computation.

Parameters:: bias – [in] flag bias.

double mean_absolute_deviation_computation(double location) const#

Computation of the mean absolute deviation.

Parameters:: location – [in] location measure (e.g. mean or median).
Returns:: mean absolute deviation.

double log_geometric_mean_computation() const#

Computation of the log of the geometric mean.

Returns:: log geometric mean.

double skewness_computation() const#

Computation of the coefficient of skewness.

Returns:: coefficient of skewness.

double kurtosis_computation() const#

Computation of the excess kurtosis: excess kurtosis = coefficient of kurtosis - 3.

Returns:: excess kurtosis.

void mean_direction_computation(double *mean_direction) const#

Computation of the mean direction for a circular variable.

Parameters:: mean_direction – [in] pointer on the mean direction.

double information_computation() const#

Computation of the information quantity.

Returns:: information quantity.

double likelihood_computation(const Distribution &dist) const#

Computation of the log-likelihood of a discrete distribution for a sample.

Parameters:: dist – [in] reference on a Distribution object.
Returns:: log-likelihood.

void distribution_estimation(Distribution *pdist) const#

Estimation of a categorical distribution on the basis of a frequency distribution.

Parameters:: dist – [in] pointer on a Distribution object.

void penalized_likelihood_estimation(Distribution *dist, double weight, penalty_type pen_type, double *penalty, side_effect outside) const#

Estimation of a discrete distribution on the basis of a frequency distribution using a penalized likelihood estimator.

Parameters:

dist – [in] pointer on a Distribution object,
weight – [in] penalty weight,
pen_type – [in] penalty type (first- or second-order difference or entropy),
penalty – [in] penalty,
outside – [in] management of side effects (zero outside the support or continuation of the distribution).

double uniform_estimation(DiscreteParametric *pdist, int min_inf_bound, bool min_inf_bound_flag) const#

Estimation of the parameters of a uniform distribution on the basis of a frequency distribution (estimation is biased)

Parameters:

dist – [in] pointer on a DiscreteParametric object,
min_inf_bound – [in] minimum lower bound of the support,
min_inf_bound_flag – [in] flag on the distribution shift.

Returns:

maximized log-likelihood.

double binomial_estimation(DiscreteParametric *pdist, int min_inf_bound, bool min_inf_bound_flag) const#

Estimation of the parameters of a shifted binomial distribution on the basis of a frequency distribution.

Parameters:

dist – [in] pointer on a DiscreteParametric object,
min_inf_bound – [in] minimum lower bound of the support,
min_inf_bound_flag – [in] flag on the distribution shift.

Returns:

maximized log-likelihood.

double poisson_estimation(DiscreteParametric *pdist, int min_inf_bound, bool min_inf_bound_flag, double cumul_threshold) const#

Estimation of the parameters of a shifted Poisson distribution on the basis of a frequency distribution.

Parameters:

dist – [in] pointer on a DiscreteParametric object,
min_inf_bound – [in] minimum lower bound of the support,
min_inf_bound_flag – [in] flag on the distribution shift,
cumul_threshold – [in] threshold on the cumulative distribution function.

Returns:

maximized log-likelihood.

double negative_binomial_estimation(DiscreteParametric *pdist, int min_inf_bound, bool min_inf_bound_flag, double cumul_threshold) const#

Estimation of the parameters of a shifted negative binomial distribution on the basis of a frequency distribution.

Parameters:

dist – [in] pointer on a DiscreteParametric object,
min_inf_bound – [in] minimum lower bound of the support,
min_inf_bound_flag – [in] flag on the distribution shift,
cumul_threshold – [in] threshold on the cumulative distribution function.

Returns:

maximized log-likelihood.

double geometric_poisson_estimation(DiscreteParametric *pdist, int min_inf_bound, bool min_inf_bound_flag, double cumul_threshold) const#

Estimation of the parameters of a shifted Poisson geometric distribution on the basis of a frequency distribution.

Parameters:

dist – [in] pointer on a DiscreteParametric object,
min_inf_bound – [in] minimum lower bound of the support,
min_inf_bound_flag – [in] flag on the distribution shift,
cumul_threshold – [in] threshold on the cumulative distribution function.

Returns:

maximized log-likelihood.

double parametric_estimation(DiscreteParametric *pdist, int min_inf_bound = 0, bool min_inf_bound_flag = true, double cumul_threshold = CUMUL_THRESHOLD, bool geometric_poisson = false) const#

Estimation of the parameters of a discrete parametric distribution (binomial, Poisson, negative binomial, Poisson geometric) on the basis of a frequency distribution.

Parameters:

dist – [in] pointer on a DiscreteParametric object,
min_inf_bound – [in] minimum lower bound of the support,
min_inf_bound_flag – [in] flag on the distribution shift,
cumul_threshold – [in] threshold on the cumulative distribution function,
geometric_poisson – [in] flag on the estimation of a Poisson geometric distribution.

Returns:

maximized log-likelihood.

double type_parametric_estimation(DiscreteParametric *pdist, int min_inf_bound = 0, bool min_inf_bound_flag = true, double cumul_threshold = CUMUL_THRESHOLD, bool geometric_poisson = false) const#

Estimation of the type and parameters of a discrete parametric distribution (binomial, Poisson, negative binomial, Poisson geometric) on the basis of a frequency distribution.

Parameters:

dist – [in] pointer on a DiscreteParametric object,
min_inf_bound – [in] minimum lower bound of the support,
min_inf_bound_flag – [in] flag on the distribution shift,
cumul_threshold – [in] threshold on the cumulative distribution function,
geometric_poisson – [in] flag on the estimation of a Poisson geometric distribution.

Returns:

maximized log-likelihood.

DiscreteParametric *type_parametric_estimation(int min_inf_bound = 0, bool min_inf_bound_flag = true, double cumul_threshold = CUMUL_THRESHOLD) const#

Estimation of the type and parameters of a discrete parametric distribution (binomial, Poisson, negative binomial) on the basis of a frequency distribution.

Parameters:

min_inf_bound – [in] minimum lower bound of the support,
min_inf_bound_flag – [in] flag on the distribution shift,
cumul_threshold – [in] threshold on the cumulative distribution function.

Returns:

discrete parametric distribution.

void equilibrium_process_combination(const Reestimation<Type> *length_bias_reestim, double imean)#

Combination of the reestimation quantities of the basis time interval distribution and the length-biased distribution (equilibrium stochastic process).

Parameters:

length_bias_reestim – [in] pointer on the reestimation quantities of the length-biased distribution,
imean – [in] distribution mean.

void equilibrium_process_estimation(const Reestimation<Type> *length_bias_reestim, Distribution *dist, double imean) const#

Estimation of a discrete distribution on the basis of the reestimation quantities of the basis time interval distribution and the length-biased distribution (equilibrium stochastic process).

Parameters:

length_bias_reestim – [in] pointer on the reestimation quantities of the length-biased distribution,
dist – [in] distribution,
imean – [in] distribution mean.

void penalized_likelihood_equilibrium_process_estimation(const Reestimation<Type> *length_bias_reestim, Distribution *dist, double imean, double weight, penalty_type pen_type, double *penalty, side_effect outside) const#

Estimation of a discrete distribution on the basis of the reestimation quantities of the basis time interval distribution and the length-biased distribution using a penalized likelihood estimator (equilibrium stochastic process).

Parameters:

length_bias_reestim – [in] pointer on the reestimation quantities of the length-biased distribution,
dist – [in] distribution,
imean – [in] distribution mean,
weight – [in] penalty weight ,
pen_type – [in] penalty type (first- or second-order difference or entropy),
penalty – [in] penalty,
outside – [in] management of side effects (zero outside the support or continuation of the distribution).

void state_occupancy_estimation(const Reestimation<Type> *final_run, Reestimation<double> *occupancy_reestim, Type *occupancy_survivor, Type *censored_occupancy_survivor, bool characteristic_computation = true)#

Estimation of a state occupancy distribution using the Kaplan-Meier estimator.

Parameters:

final_run – [in] pointer on the right-censored sojourn times,
occupancy_reestim – [in] pointer on the reestimation quantities,
occupancy_survivor – [in] pointer on the survival function corresponding to the complete sojourn times,
censored_occupancy_survivor – [in] pointer on the survival function corresponding to the right-censored sojourn times,
characteristic_computation – [in] flag for the computation of the characteristics of the reestimation quantities.

void gamma_estimation(ContinuousParametric *dist, int iter) const#

Estimation of a gamma distribution on the basis of a frequency distribution.

Parameters:

dist – [in] continuous distribution,
iter – [in] EM iteration.

void zero_inflated_gamma_estimation(ContinuousParametric *dist, int iter) const#

Estimation of a zero-inflated gamma distribution on the basis of a frequency distribution.

Parameters:

dist – [in] continuous distribution,
iter – [in] EM iteration.

void inverse_gaussian_estimation(ContinuousParametric *dist) const#

Estimation of an inverse Gaussian distribution on the basis of a frequency distribution.

Parameters:: dist – [in] continuous distribution.

Public Members

int nb_value#: number of values from 0

int alloc_nb_value#: number of allocated values

int offset#: number of values of null frequencies from 0

Type nb_element#: sample size

Type max#: maximum frequency

double mean#: mean

double variance#: variance

Type *frequency#: frequency for each value

class Regression : public stat_tool::StatInterface, public stat_tool::RegressionKernel #

Regression function.

Public Functions

Regression()#: Default constructor of the Regression class.

Regression(parametric_function iident, int explanatory_variable, int response_variable, const Vectors &vec)#

Construction of a Regression object from a Vectors object.

Parameters:

iident – [in] regression function identifier,
explanatory_variable – [in] explanatory variable index,
response_variable – [in] response variable index,
vec – [in] reference on a Vectors object.

Regression(const Regression &regression)#

Constructor by copy of the Regression class.

Parameters:: regression – [in] reference on a Regression object.

~Regression()#: Destructor of the Regression class.

Regression &operator=(const Regression &regression)#

Assignment operator of the Regression class.

Parameters:: regression – [in] reference on a Regression object.
Returns:: Regression object.

virtual std::ostream &line_write(std::ostream &os) const#

Writing on a single line of a Regression object.

Parameters:: os – [inout] stream.

virtual bool spreadsheet_write(StatError &error, const std::string path) const#

Writing of a Regression object in a file at the spreadsheet format.

Parameters:

error – [in] reference on a StatError object,
path – [in] file path.

Returns:

error status.

virtual bool plot_write(StatError &error, const char *prefix, const char *title = NULL) const#

Plot of a Regression object using Gnuplot.

Parameters:

error – [in] reference on a StatError object,
prefix – [in] file prefix,
title – [in] figure title.

Returns:

error status.

virtual MultiPlotSet *get_plotable() const#

Plot of a Regression object.

Returns:: MultiPlotSet object.

class RegressionKernel#

Regression kernel class.

Subclassed by stat_tool::Regression

Public Functions

void copy(const RegressionKernel&)#

Copy of a RegressionKernel object.

Parameters:: regression – [in] reference on a RegressionKernel object.

void remove()#: Destruction of the data members of a RegressionKernel object.

RegressionKernel()#: Default constructor of the RegressionKernel class.

RegressionKernel(parametric_function iident, int imin_value, int imax_value)#

Constructor of the RegressionKernel class.

Parameters:

iident – [in] identifier,
imin_value – [in] minimum value of the explanatory variable,
imax_value – [in] maximum value of the explanatory variable.

~RegressionKernel()#: Destructor of the RegressionKernel class.

RegressionKernel &operator=(const RegressionKernel &regression)#

Assignment operator of the RegressionKernel class.

Parameters:: regression – [in] reference on a RegressionKernel object.
Returns:: RegressionKernel object.

std::ostream &ascii_parameter_print(std::ostream &os) const#

Writing of the regression function parameters.

Parameters:: os – [inout] stream.

std::ostream &ascii_formal_print(std::ostream &os) const#

Formal writing of the regression function.

Parameters:: os – [inout] stream.

std::ostream &ascii_print(std::ostream &os) const#

Writing of the regression function.

Parameters:: os – [inout] stream.

std::ostream &spreadsheet_print(std::ostream &os) const#

Writing of the regression function at the spreadsheet format.

Parameters:: os – [inout] stream.

bool plot_print(const char *path) const#

Writing of the regression function (Gnuplot output).

Parameters:: path – [in] file path.
Returns:: error status.

void plotable_write(SinglePlot &plot) const#

Writing of the regression function for plotting.

Parameters:: plot – [in] reference on a SinglePlot object.

void computation()#: Computation of a parametric regression function.

double min_computation() const#

Computation of the minimum value of a regression function.

Returns:: minimum value.

double max_computation() const#

Computation of the maximum value of a regression function.

Returns:: maximum value.

Public Members

parametric_function ident#: identifier of the regression function

int min_value#: minimum value

int max_value#: maximum value

double regression_df#: degrees of freedom regression

double residual_df#: degrees of freedom residuals

int nb_parameter#: number of parameters

double *parameter#: parameters

double *point#: points

class SinglePlot#

class StatError#

Class for error management.

Public Functions

StatError(int imax_nb_error = NB_ERROR)#

Constructor of the StatError class.

Parameters:: imax_nb_error – [in] maximum number of errors.

~StatError()#: Destructor of the StatError class.

void update(const char *ilabel, int iline = 0, int icolumn = 0)#

Update of a StatError object.

Parameters:

ilabel – [in] label,
iline – [in] line,
icolumn – [in] column.

void correction_update(const char *ilabel, const char *correction, int iline = 0, int icolumn = 0)#

Update of a StatError object.

Parameters:

ilabel – [in] label,
correction – [in] correction,
iline – [in] line,
icolumn – [in] column.

void correction_update(const char *ilabel, int correction, int iline = 0, int icolumn = 0)#

Update of a StatError object.

Parameters:

ilabel – [in] label,
correction – [in] correction,
iline – [in] line,
icolumn – [in] column.

class StatInterface#

Abstract class defining a common interface.

Subclassed by stat_tool::Compound, stat_tool::CompoundData, stat_tool::Convolution, stat_tool::ConvolutionData, stat_tool::Dendrogram, stat_tool::DiscreteDistributionData, stat_tool::DiscreteMixture, stat_tool::DiscreteMixtureData, stat_tool::DiscreteParametricModel, stat_tool::DistanceMatrix, stat_tool::MultivariateMixture, stat_tool::Regression, stat_tool::VectorDistance, stat_tool::Vectors

Public Functions

std::string ascii_write(bool exhaustive = false) const#

Writing of a StructureAnalysis object (for display on the console).

Parameters:: exhaustive – [in] flag detail level,
Returns:: string.

template<typename Type> class TemplateMultiPlotSet#

Public Members

std::vector<int> variable#: identifier / number of the variable to be displayed

std::vector<Type> viewpoint#: nature of the graph, for example observation distribution…

class Test#

Test of hypothesis.

Public Functions

void copy(const Test &test)#

Construction by copy of a Test object.

Parameters:: test – [in] reference on a Test object.

Test(test_distribution iident, bool ione_side = true)#

Constructor of the Test class.

Parameters:

iident – [in] identifier,
ione_side – [in] flag one-sided/two-sided.

Test(test_distribution iident, bool ione_side, int idf1, int idf2, double ivalue)#

Constructor of the Test class.

Parameters:

iident – [in] identifier,
ione_side – [in] flag one-sided/two-sided,
idf1 – [in] degrees of freedom,
idf2 – [in] degrees of freedom,
ivalue – [in] value.

Test(test_distribution iident, bool ione_side, int idf1, int idf2, double ivalue, double icritical_probability)#

Constructor of the Test class.

Parameters:

iident – [in] identifier,
ione_side – [in] flag one-sided/two-sided,
idf1 – [in] degrees of freedom,
idf2 – [in] degrees of freedom,
ivalue – [in] value,
icritical_probability – [in] critical probability.

Test &operator=(const Test &test)#

Assignment operator of the Test class.

Parameters:: test – [in] reference on a Test object.
Returns:: Test object.

std::ostream &ascii_print(std::ostream &os, bool comment_flag = false, bool reference_flag = true) const#

Writing of the results of a test.

Parameters:

os – [inout] stream,
comment_flag – [in] flag comment,
reference_flag – [in] flag reference result.

std::ostream &spreadsheet_print(std::ostream &os, bool reference_flag = true) const#

Writing of the results of a test at the spreadsheet format.

Parameters:

os – [inout] stream,
reference_flag – [in] flag reference result.

void standard_normal_critical_probability_computation()#: Computation of the critical probability from the value taken by a standard Gaussian random variable.

void standard_normal_value_computation()#: Computation of the value taken by a standard Gaussian random variable from the critical probability.

void chi2_critical_probability_computation()#: Computation of la critical probability from the value taken by a Chi2 random variable.

void chi2_value_computation()#: Computation of the value taken by a Chi2 random variable from the critical probability.

void F_critical_probability_computation()#: Computation of the critical probability from the value taken by a F random variable.

void F_value_computation()#: Computation of the value taken by a F random variable from the critical probability.

void t_critical_probability_computation()#: Computation of the critical probability from the value taken by a Student’s t-random variable.

void t_value_computation()#: Computation of the value taken by a Student’s t-random variable from the critical probability.

Public Members

test_distribution ident#: identifier (Normal/Chi2/F/Student’s t)

bool one_side#: one-sided/two-sided

int df1#: degrees of freedom (Chi2/F/Student’s t)

int df2#: degrees of freedom (F)

double value#: statistic value

double critical_probability#: critical probability

class VectorDistance : public stat_tool::StatInterface #

Parameterization of a distance between vectors with heterogeneous variables.

Public Functions

VectorDistance()#: Default constructor of the VectorDistance class.

VectorDistance(int inb_variable, variable_type *ivar_type, double *iweight, metric idistance_type = ABSOLUTE_VALUE)#

Constructor of the VectorDistance class.

Parameters:

inb_variable – [in] number of variables,
ivar_type – [in] variable types,
iweight – [in] variable weights,
idistance_type – [in] distance type (ABSOLUTE_VALUE/QUADRATIC).

VectorDistance(int inb_variable, metric idistance_type, variable_type *ivar_type, double *iweight, int *inb_value, double ***icategory_distance, int *iperiod)#

Constructor of the VectorDistance class.

Parameters:

inb_variable – [in] number of variables,
idistance_type – [in] distance type (ABSOLUTE_VALUE/QUADRATIC),
ivar_type – [in] variable types,
iweight – [in] variable weights,
inb_value – [in] number of categories (for categorical variables),
icategory_distance – [in] between-category distance matrices (for categorical variables),
iperiod – [in] periods (for circular variables).

~VectorDistance()#: Destructor of the VectorDistance class.

VectorDistance &operator=(const VectorDistance &vector_dist)#

Assignment operator of the VectorDistance class.

Parameters:: vector_dist – [in] reference on a VectorDistance object.
Returns:: VectorDistance object.

virtual std::ostream &line_write(std::ostream &os) const#

Writing on a single line of a VectorDistance object.

Parameters:: os – [inout] stream.

double *max_category_distance_computation(int variable) const#

Computation of the maximum distance for each category.

Parameters:: variable – [in] variable index.
Returns:: maximum distances.

void dispersion_update(int variable, double idispersion) const#

Update of the standardization quantity.

Parameters:

variable – [in] variable index,
idispersion – [in] variable dispersion.

void dispersion_computation(int variable, const FrequencyDistribution *marginal_distribution, double *rank = NULL) const#

Computation of the standardization quantity.

Parameters:

variable – [in] variable index,
marginal_distribution – [in] marginal frequency distribution,
rank – [in] ranks.

Public Static Functions

static VectorDistance *ascii_read(StatError &error, const std::string path)#

Construction of a VectorDistance object from a file.

Parameters:

error – [in] reference on a StatError object,
path – [in] file path.

Returns:

VectorDistance object.

class Vectors : public stat_tool::StatInterface #

Vectors with integer- and real-valued variables.

Subclassed by stat_tool::MultivariateMixtureData

Public Functions

Vectors()#: Default constructor of the Vectors class.

Vectors(int inb_vector, int *iidentifier, int inb_variable, int **iint_vector)#

Constructor of the Vectors class.

Parameters:

inb_vector – [in] number of individuals,
iidentifier – [in] individual identifiers,
inb_variable – [in] number of variables,
iint_vector – [in] integer-valued vectors.

Vectors(int inb_vector, int *iidentifier, int inb_variable, double **ireal_vector)#

Constructor of the Vectors class.

Parameters:

inb_vector – [in] number of individuals,
iidentifier – [in] individual identifiers,
inb_variable – [in] number of variables,
ireal_vector – [in] real-valued vectors.

Vectors(int inb_vector, int *iidentifier, int inb_variable, variable_nature *itype, int **iint_vector, double **ireal_vector, bool variable_index = true)#

Constructor of the Vectors class.

Parameters:

inb_vector – [in] number of individuals,
iidentifier – [in] individual identifiers,
inb_variable – [in] number of variables,
itype – [in] variable types,
iint_vector – [in] integer variables,
ireal_vector – [in] real variables,
variable_index – [in] variable indexing.

Vectors(const Vectors &vec, int variable, variable_nature itype)#

Constructor of the Vectors class.

Parameters:

vec – [in] reference on a Vectors object,
variable – [in] variable index,
itype – [in] variable type.

Vectors(const Vectors &vec, int inb_vector, int *index)#

Constructor of the Vectors class.

Parameters:

vec – [in] reference on a Vectors object,
inb_vector – [in] number of individuals,
index – [in] indices of the selected individuals.

Vectors(const Vectors &vec, vector_transformation transform = VECTOR_COPY)#

Constructor by copy of the Vectors class.

Parameters:

vec – [in] reference on a Vectors object,
transform – [in] type of transform.

~Vectors()#: Destructor of the Vectors class.

Vectors &operator=(const Vectors &vec)#

Assignment operator of the Vectors class.

Parameters:: vec – [in] reference on a Vectors object.
Returns:: Vectors object.

void min_interval_computation(int variable)#

Computation of the shortest interval between 2 successive values for a variable.

Parameters:: variable – [in] variable index.

DiscreteDistributionData *extract(StatError &error, int variable) const#

Extraction of the marginal frequency distribution for a positive integer-valued variable.

Parameters:

error – [in] reference on a StatError object,
variable – [in] variable index.

Returns:

DiscreteDistributionData object.

bool check(StatError &error)#

Checking of the identifiers of a Vectors object.

Parameters:: error – [in] reference on a StatError object.
Returns:: error status.

Vectors *merge(StatError &error, int nb_sample, const Vectors **ivec) const#

Merging of Vectors objects.

Parameters:

error – [in] reference on a StatError object,
nb_sample – [in] number of Vectors objects,
ivec – [in] pointer on the Vectors objects.

Returns:

Vectors object.

Vectors *shift(StatError &error, int variable, int shift_param) const#

Shifting of values of a variable.

Parameters:

error – [in] reference on a StatError object,
variable – [in] variable index,
shift_param – [in] integer shifting parameter.

Returns:

Vectors object.

Vectors *shift(StatError &error, int variable, double shift_param) const#

Shifting of values of a real-valued variable.

Parameters:

error – [in] reference on a StatError object,
variable – [in] variable index,
shift_param – [in] real shifting parameter.

Returns:

Vectors object.

Vectors *thresholding(StatError &error, int variable, int threshold, threshold_direction mode) const#

Thresholding of values of a variable.

Parameters:

error – [in] reference on a StatError object,
variable – [in] variable index,
threshold – [in] integer threshold,
mode – [in] mode (ABOVE/BELOW).

Returns:

Vectors object.

Vectors *thresholding(StatError &error, int variable, double threshold, threshold_direction mode) const#

Thresholding of values of a real-valued variable.

Parameters:

error – [in] reference on a StatError object,
variable – [in] variable index,
threshold – [in] real threshold,
mode – [in] mode (ABOVE/BELOW).

Returns:

Vectors object.

Vectors *transcode(StatError &error, int variable, int *category) const#

Transcoding of categories of an integer-valued variable.

Parameters:

error – [in] reference on a StatError object,
variable – [in] variable index,
category – [in] transcoding table.

Returns:

Vectors object.

Vectors *cluster(StatError &error, int variable, int step, rounding mode = FLOOR) const#

Clustering of values of a variable.

Parameters:

error – [in] reference on a StatError object,
variable – [in] variable index,
step – [in] clustering step,
mode – [in] mode (FLOOR/ROUND/CEIL).

Returns:

Vectors object.

Vectors *cluster(StatError &error, int variable, int inb_value, int *ilimit) const#

Partitioning of values of a variable.

Parameters:

error – [in] reference on a StatError object,
variable – [in] variable index,
nb_class – [in] number of classes,
ilimit – [in] integer limits between classes (beginning of classes).

Returns:

Vectors object.

Vectors *cluster(StatError &error, int variable, int nb_class, double *ilimit) const#

Partitioning of values of a real-valued variable.

Parameters:

error – [in] reference on a StatError object,
variable – [in] variable index,
nb_class – [in] number of classes,
ilimit – [in] real limits between classes (beginning of classes).

Returns:

Vectors object.

Vectors *scaling(StatError &error, int variable, int scaling_coeff) const#

Scaling of a variable.

Parameters:

error – [in] reference on a StatError object,
variable – [in] variable index,
scaling_coeff – [in] integer scaling factor.

Returns:

Vectors object.

Vectors *scaling(StatError &error, int variable, double scaling_coeff) const#

Scaling of a variable.

Parameters:

error – [in] reference on a StatError object,
variable – [in] variable index,
scaling_coeff – [in] real scaling factor.

Returns:

Vectors object.

Vectors *round(StatError &error, int variable = I_DEFAULT, rounding mode = ROUND) const#

Rounding of values of a real-valued variable.

Parameters:

error – [in] reference on a StatError object,
variable – [in] variable index,
mode – [in] mode (FLOOR/ROUND/CEIL).

Returns:

Vectors object.

Vectors *log_transform(StatError &error, int variable = I_DEFAULT, log_base base = NATURAL) const#

Log-transform of values.

Parameters:

error – [in] reference on a StatError object,
variable – [in] variable index,
base – [in] base of the logarithm.

Returns:

Vectors object.

Vectors *select_individual(StatError &error, int inb_vector, int *iidentifier, bool keep = true) const#

Selection of individuals by their identifiers.

Parameters:

error – [in] reference on a StatError object,
inb_vector – [in] number of individuals,
iidentifier – [in] identifiers,
keep – [in] flag for keeping or rejecting the selected individuals.

Returns:

Vectors object.

Vectors *select_variable(StatError &error, int inb_variable, int *ivariable, bool keep = true) const#

Selection of variables.

Parameters:

error – [in] reference on a StatError object,
inb_variable – [in] number of variables,
ivariable – [in] variable indices,
keep – [in] flag for keeping or rejecting the selected variables.

Returns:

Vectors object.

Vectors *sum_variable(StatError &error, int nb_summed_variable, int *ivariable) const#

Summation of variables.

Parameters:

error – [in] reference on a StatError object,
nb_summed_variable – [in] number of variables to be summed,
ivariable – [in] variable indices.

Returns:

Vectors object.

Vectors *merge_variable(StatError &error, int nb_sample, const Vectors **ivec, int ref_sample = I_DEFAULT) const#

Merging of variables of Vectors objects.

Parameters:

error – [in] reference on a StatError object,
nb_sample – [in] number of Vectors objects,
ivec – [in] pointer on the Vectors objects,
ref_sample – [in] reference Vectors object for the identifiers.

Returns:

Vectors object.

virtual std::ostream &line_write(std::ostream &os) const#

Writing on a single line of a Vectors object.

Parameters:: os – [inout] stream.

std::string ascii_data_write(bool exhaustive = false) const#

Writing of a Vectors object.

Parameters:: exhaustive – [in] flag detail level,
Returns:: string.

virtual bool spreadsheet_write(StatError &error, const std::string path) const#

Writing of a Vectors object in a file at the spreadsheet format.

Parameters:

error – [in] reference on a StatError object,
path – [in] file path.

Returns:

error status.

virtual bool plot_write(StatError &error, const char *prefix, const char *title = NULL) const#

Plot of a of a Vectors object using Gnuplot.

Parameters:

error – [in] reference on a StatError object,
prefix – [in] file prefix,
title – [in] figure title.

Returns:

error status.

virtual MultiPlotSet *get_plotable() const#

Plot of a Vectors object.

Returns:: MultiPlotSet object.

bool select_bin_width(StatError &error, int variable, double bin_width, double imin_value = D_INF)#

Change of the bin width of the marginal histogram for a variable.

Parameters:

error – [in] reference on a StatError object,
variable – [in] variable index,
bin_width – [in] bin width,
imin_value – [in] minimum value,

Returns:

error status.

int cumulative_distribution_function_computation(int variable, double **cdf) const#

Computation of the cumulative frequency distribution function for a variable.

Parameters:

variable – [in] variable index,
cdf – [in] (values, cumulative distribution function).

Returns:

cumulative frequency distribution function.

double mean_absolute_deviation_computation(int variable, double location) const#

Computation of the mean absolute deviation for a variable.

Parameters:

variable – [in] variable index,
location – [in] location measure (e.g. mean or median).

Returns:

mean absolute deviation.

double mean_absolute_difference_computation(int variable) const#

Computation of the mean absolute difference for a variable.

Parameters:: variable – [in] variable index.
Returns:: mean absolute difference.

double skewness_computation(int variable) const#

Computation of the coefficient of skewness for a variable.

Parameters:: variable – [in] variable index.
Returns:: coefficient of skewness.

double kurtosis_computation(int variable) const#

Computation of the excess kurtosis for a variable: excess kurtosis = coefficient of kurtosis - 3.

Parameters:: variable – [in] variable index.
Returns:: excess kurtosis.

double *mean_direction_computation(int variable, angle_unit unit) const#

Computation of the mean direction for a circular variable.

Parameters:

variable – [in] variable index,
unit – [in] unit (DEGREE/RADIAN).

Returns:

mean direction.

double spearman_rank_single_correlation_computation() const#

Computation of the Spearman rank correlation coefficient between 2 variables.

Returns:: correlation coefficient.

double kendall_rank_single_correlation_computation() const#

Computation of the Kendall rank correlation coefficient between 2 variables.

Returns:: correlation coefficient.

bool rank_correlation_computation(StatError &error, std::ostream *os, correlation_type correl_type, const std::string path = "") const#

Computation of a rank correlation coefficient matrix (either in the Spearman or in the Kendall sense).

Parameters:

error – [in] reference on a StatError object,
os – [in] stream for displaying the rank correlation coefficient matrix,
correl_type – [in] rank correlation coefficient type (SPEARMAN/KENDALL),
path – [in] file path.

Returns:

error status.

DistanceMatrix *comparison(StatError &error, const VectorDistance &ivector_dist, bool standardization = true) const#

Comparison of vectors (computation of the matrix of pairwise distances between vectors).

Parameters:

error – [in] reference on a StatError object,
ivector_dist – [in] reference on a VectorDistance object,
standardization – [in] flag standardization (only for variables of the same type).

Returns:

DistanceMatrix object.

bool contingency_table(StatError &error, std::ostream *os, int variable1, int variable2, const std::string path = "", output_format format = ASCII) const#

Computation of a contingency table for 2 categorical variables.

Parameters:

error – [in] reference on a StatError object,
os – [in] stream for displaying the contingency table,
variable1 – [in] variable 1 index,
variable2 – [in] variable 2 index,
path – [in] file path,
format – [in] file format (ASCII/SPREADSHEET).

Returns:

error status.

bool variance_analysis(StatError &error, std::ostream *os, int class_variable, int response_variable, int response_type, const std::string path = "", output_format format = ASCII) const#

One-way analysis of variance.

Parameters:

error – [in] reference on a StatError object,
os – [in] stream for displaying the ANOVA results,
class_variable – [in] explanatory variable index,
response_variable – [in] response variable index,
response_type – [in] response variable type (ORDINAL/NUMERIC),
path – [in] file path,
format – [in] file format (ASCII/SPREADSHEET).

Returns:

error status.

bool sup_norm_distance(StatError &error, std::ostream &os, const Vectors &ivec) const#

Computation of sup norm distance between two empirical continuous distributions.

Parameters:

error – [in] reference on a StatError object,
os – [in] stream for displaying the sup norm distance,
ivec – [in] reference on a Vector object.

Returns:

error status.

Regression *linear_regression(StatError &error, int explanatory_variable, int response_variable) const#

Linear regression.

Parameters:

error – [in] reference on a StatError object,
explanatory_variable – [in] explanatory variable index,
response_variable – [in] response variable index.

Returns:

Regression object.

Regression *moving_average(StatError &error, int explanatory_variable, int response_variable, int nb_point, double *filter, moving_average_method algorithm = AVERAGING) const#

Nonparametric moving-average regression.

Parameters:

error – [in] reference on a StatError object,
explanatory_variable – [in] explanatory variable index,
response_variable – [in] response variable index,
nb_point – [in] filter half width,
filter – [in] filter,
algorithm – [in] response computation (AVERAGING/LEAST_SQUARES).

Returns:

Regression object.

Regression *moving_average(StatError &error, int explanatory_variable, int response_variable, const Distribution &dist, moving_average_method algorithm = AVERAGING) const#

Nonparametric moving-average regression.

Parameters:

error – [in] reference on a StatError object,
explanatory_variable – [in] explanatory variable index,
response_variable – [in] response variable index,
dist – [in] symmetric distribution,
algorithm – [in] response computation (AVERAGING/LEAST_SQUARES).

Returns:

Regression object.

Regression *nearest_neighbor_smoother(StatError &error, int explanatory_variable, int response_variable, double span, bool weighting = true) const#

Nonparametric k-nearest-neighbor regression.

Parameters:

error – [in] reference on a StatError object,
explanatory_variable – [in] explanatory variable index,
response_variable – [in] response variable index,
span – [in] neighboring proportion with respect to the sample size,
weighting – [in] flag neighbor weighting.

Returns:

Regression object.

inline FrequencyDistribution *get_marginal_distribution(int variable) const#: get marginal counts for a given variable (does not allocate a new object, returns internal pointer)

inline Histogram *get_marginal_histogram(int variable) const#: get marginal histogram for a given variable (does not allocate a new object, returns internal pointer)

Public Static Functions

static Vectors *build(StatError &error, const std::vector<std::vector<int>> &iint_vector, const std::vector<std::vector<double>> &ireal_vector, const std::vector<int> &iidentifier)#

Construction of a Vectors objects from arrays of discrete values, real values and identifiers.

Parameters:

error – [in] reference on a StatError object,
iint_vector – [in] integer variables,
ireal_vector – [in] real variables,
iidentifier – [in] individual identifiers.

static Vectors *ascii_read(StatError &error, const std::string path)#

Construction of a Vectors object from a file. Format: rows: individuals | columns: variables.

Parameters:

error – [in] reference on a StatError object,
path – [in] file path.

Returns:

Vectors object.