gtm_demo {gtm} | R Documentation |
Demonstrates the GTM with a 2D target space and a 1D latent space
Description
This script generates a simple data set in 2 dimensions, with an intrinsic dimensionality of 1, and trains a GTM with a 1-dimensional latent variable to model this data set, visually illustrating the training process.
Usage
gtm_demo(randomSetup = FALSE, points = 5, samples = 20)
Arguments
randomSetup |
if FALSE, use gtm_pci_beta for initial configuration, otherwise, initialize using gtm_ri |
points |
number of Gaussians |
samples |
number of latent space samples |
Value
List with elements
X
|
latent sample points |
T
|
data in the target space |
W
|
the initialised weight matrix |
Y
|
centers of Gaussians |
beta
|
the optimized beta value, scalar. |
llh
|
log-likelihood |
FI
|
basis functions' activation |
See Also
gtm_ri
, gtm_pci
Examples
gtm_demo ()
gtm_demo (randomSetup = TRUE)
[Package
gtm version 1.0
Index]