GLLiM

GLLiM class overall

class GLLiM(L, D, K, gamma_type, sigma_type)

Gaussian Locally-Linear Model (GLLiM) for probabilistic modeling.

Parameters:

• L (int) – The latent space dimension.

• D (int) – The observed space dimension.

• K (int) – The number of Gaussian components.

• gamma_type (str) – The type of gamma parameter among {‘full’, ‘diag’, ‘iso’}.

• sigma_type (str) – The type of sigma parameter among {‘full’, ‘diag’, ‘iso’}.

Returns:

An instance of the GLLiM class.

Main methods

initialize	Initialize the GLLiM model with given data and parameters.
train	Train the GLLiM model with given data and parameters.
getInverse	Get the inverse parameters of the GLLiM model.
directDensities	Compute the direct densities given input matrix x and its uncertainties.
inverseDensities	Compute the inverse densities given input matrix y and its uncertainties.
getInsights	Returns ann Insights structure with informations about initialisation and training time, log-likelihood and arguments.

Getters

getDimensions	Get the dimensions of the GLLiM model.
getConstraints	Get the constraints of the GLLiM model.
getParams	Get the parameters of the GLLiM model.
getParamPi	Get the mixture coefficients Pi.
getParamA	Get the parameter matrix A.
getParamB	Get the parameter matrix B.
getParamC	Get the parameter matrix C.
getParamGamma	Get the gamma parameters.
getParamSigma	Get the sigma parameters.

Setters

setParams	Set the parameters of the GLLiM model.
setParamPi	Set the mixture coefficients Pi.
setParamA	Set the parameter matrix A.
setParamB	Set the parameter matrix B.
setParamC	Set the parameter matrix C.
setParamGamma	Set the gamma parameters.
setParamSigma	Set the sigma parameters.

Structures

GLLiMParameters	Describes the parameters of the GLLiM model theta = {Pi, A, B, C, Gamma, Sigma}.
GLLiMConstraints	Describes the constraints of the covariance matrices Gamma and Sigma.
PredictionResult	Describes the results concerning a GLLiM density estimation (direct or inverse).
FullGMMResult	Describes the results concerning a GLLiM density estimation by the mean.
MergedGMMResult	Describes the results concerning a GLLiM density estimation by the centroids.
Insights	Describes valuable information about initialisation and training (time, log-likelihood and configuration).
InitialisationInsights	Describes valuable information about initialisation.
TrainingInsights	Describes valuable information about training.