gtm_ppd2 {gtm} | R Documentation |
Latent space posterior probability distribution for a given data point
Description
This function calculates the posterior probability distribution induced in the latent space of a trained GTM model for a given data point, and returns it in a format suitable for 3D graphic plotting routines
Usage
gtm_ppd2(t, Y, beta, X, xDim, yDim)
Arguments
t |
a point in the data space; 1-by-D |
Y |
centres of the Gaussian mixture generated by the GTM in the data space, Y = FI*W; K-by-D |
beta |
variance of Gaussian mixture; scalar |
X |
latent sample |
xDim,yDim |
number of points along the 2 dimensions of the latent space meshgrid sample |
Value
List with components
X,Y |
latent sample; if the latent space is 2D, X and Y
are mesh matrices |
Z |
posterior distribution over latent space given data point t |
See Also
gtm_pmn
, gtm_pmd
[Package
gtm version 1.0
Index]