The GTM Toolbox

gtm_bi

beta initialisation - calculate an initial value for beta

gtm_demo

demonstrates the GTM with a 2D target space and a 1D latent space

gtm_dist

distances - calculate the squared distances between two sets of data points

gtm_dstg

distances - calculate the squared distances between two sets of data points; uses global variables

gtm_gbf

Gaussian basis functions - calculates the output of Gaussian basis functions for a given set of input

gtm_hxg

hexagonal grid - produces a 2D grid with points arranged in a hexagonal lattice

gtm_lbf

linear basis functions - calculates the output of linear basis functions for a given set of input

gtm_m2r

mesh to rows - converts from a mesh-matrix to vector representation

gtm_pca

principal components analysis - calculates the principal components of a data set

gtm_pci

principal components initialisation - returns a weight matrix initialised using principal components

gtm_pmd

posterior mode - calculates the posterior mode projection of data into the latent space

gtm_pmn

posterior mean - calculates the posterior mean projection of data into the latent space

gtm_ppd

posterior probability distribution - posterior distribution over the latent space posterior for a given data point

gtm_r2m

rows to mesh - converts data from column vector to mesh-matrix representation

gtm_rctg

rectangular grid - produces a 2D grid with points arranged in a rectangular lattice

gtm_resp

responsabilities - calculate log-likelihood and component responsabilities over a Gaussian mixture

gtm_ri

random initialisation - returns an initial random weight matrix

gtm_rspg

responsabilities - calculate log-likelihood and component responsabilities over a Gaussian mixture; uses global variables

gtm_sort

sorts the columns of argument matrix R in increasing order

gtm_stp1

setup 1D - generates the components of a GTM with a 1-dimensional latent space

gtm_stp2

setup 2D - generates the components of a GTM with a 2-dimensional latent space

gtm_trn

train - optimize (train) the parameters of a GTM model, using an EM algorithm