wolffd@0: wolffd@0: wolffd@0: wolffd@0: Netlab Reference Manual demmdn1 wolffd@0: wolffd@0: wolffd@0: wolffd@0:

demmdn1 wolffd@0:

wolffd@0:

wolffd@0: Purpose wolffd@0:

wolffd@0: Demonstrate fitting a multi-valued function using a Mixture Density Network. wolffd@0: wolffd@0:

wolffd@0: Synopsis wolffd@0:

wolffd@0:

wolffd@0: demmdn1

wolffd@0: wolffd@0: wolffd@0:

wolffd@0: Description wolffd@0:

wolffd@0: The problem consists of one input variable wolffd@0: x and one target variable t with data generated by wolffd@0: sampling t at equal intervals and then generating target data by wolffd@0: computing t + 0.3*sin(2*pi*t) and adding Gaussian noise. A wolffd@0: Mixture Density Network with 3 centres in the mixture model is trained wolffd@0: by minimizing a negative log likelihood error function using the scaled wolffd@0: conjugate gradient optimizer. wolffd@0: wolffd@0:

The conditional means, mixing coefficients and variances are plotted wolffd@0: as a function of x, and a contour plot of the full conditional wolffd@0: density is also generated. wolffd@0: wolffd@0:

wolffd@0: See Also wolffd@0:

wolffd@0: mdn, mdnerr, mdngrad, scg

wolffd@0: Pages: wolffd@0: Index wolffd@0:

wolffd@0: