gtm_bi {gtm}R Documentation

Calculate an initial value for beta

Description

The value is calculated from the average distance between the nearest neighbours in Y, the centres of the constrained Gaussian mixture generated in the target space from latent sample.

Usage

beta = gtm_bi(Y)

Arguments

Y a matrix containing the positions of the centres of the Gaussian mixture induced in target space from the latent variable samples.

Value

beta - an initial value for the inverse variance of the Gaussian mixture

Examples

  T = matrix(3:61/20, ncol=1);
  T = cbind(T, T + 1.25 * sin(2*T));
  # plot of sample data 
  plot(T[,1], T[,2], col='red', pch=21, xlim=c(0, 3.5), ylim=c(0,3.5))
    
  # setup and training
  X = gtm_pts(20)
  MU = gtm_pts(5)
  sigma = 1
  FI = gtm_gbf(MU, sigma, X)
  W = gtm_ri(T, FI);
  Y = FI 
  
  # plot of initialized data
  matplot(Y[,1], Y[,2], col='green', add=TRUE, type='p', pch=23)
  beta = gtm_bi(Y)
  symbols(x=Y[,1], y=Y[,2], circles=matrix(sqrt(1/beta) ,ncol=1,nrow=nrow(Y)),
     fg="green",add=TRUE, inches=FALSE);
  

[Package gtm version 1.0 Index]