Pseudo-density Estimation with One-Class Gaussian Process Classifers
김현철 · 한국정보과학회 학술발표논문집 · 2008
Gaussian process classifiers (GPCs) are fully statistical kernel classification models which have a latent function with Gaussian process prior. We propose a pseudo-density estimation method based on a latent function of a one-class Gaussian process classifer. Through simulation we show that a latent function of a one-class Gaussian process classifier represents topology of the pseudo-density This pseudo-density could be applied to clustering.