Title :
Simple and Effective Probability Density Estimation and Classification
Author_Institution :
Inst. of Comput. Sci., Czech Acad. of Sci., Prague
Abstract :
A mapping is introduced which maps the true distribution to a function of distances. The approximation of this function in the form of a suitable power of the distance is presented. The exponent in this approximation is used for probability density estimation in high-dimensional spaces and for classification
Keywords :
Bayes methods; estimation theory; pattern classification; probability; Bayes method; function power approximation; multivariate data classification; probability density estimation; probability distribution mapping function; Bayesian methods; Chromium; Computer science; Decision trees; Electronic mail; Nearest neighbor searches; Neural networks; Probability distribution; Tin; Classification; Multivariate data; Power approximation; Probability distribution mapping function;
Conference_Titel :
SICE-ICASE, 2006. International Joint Conference
Conference_Location :
Busan
Print_ISBN :
89-950038-4-7
Electronic_ISBN :
89-950038-5-5
DOI :
10.1109/SICE.2006.315031
