demo_multiclass_KDE

Contents

Description of demo_multiclass_KDE.m
usage of generative Gaussian model
usage of generative RBF kernel density estimation model
usage of generative polynomial kernel density estimation model

Description of demo_multiclass_KDE.m

Demonstrates generative kernel density estimation with RBF and polynomial kernels, with a Gaussian maximum likelihood fit as baseline

clear all
close all

generateData_clustersXonly

usage of generative Gaussian model

options_gs = [];
model_gs = ml_generative_Gaussian(Xtrain, ytrain, options_gs);
figure;
plotPDF(Xtrain, model_gs);
title('Generative Gaussian Model');

usage of generative RBF kernel density estimation model

options_kde = [];
options_kde.kernelOptions = struct('sigma',.75);
model_kde = ml_generative_KDE(Xtrain, ytrain, options_kde);
figure;
title('Generative RBF Kernel Density Estimation Model');
plotPDF(Xtrain, model_kde);

usage of generative polynomial kernel density estimation model

options_kde_poly = [];
options_kde_poly.kernelOptions = struct('order',3,'bias',1);
options_kde_poly.kernelOptions.kernelFunc = @ml_kernel_poly;
model_kde_poly = ml_generative_KDE(Xtrain, ytrain, options_kde_poly);
figure;
title('Generative Poly Kernel Density Estimation Model');
plotPDF(Xtrain, model_kde_poly);