Back to Search Start Over

The Effect Evaluation of Density Estimation through Non-Gaussian Measurement

Authors :
Zhiyong An
Hailan Jiang
Feng Zhao
Source :
Applied Mathematics & Information Sciences. 8:761-764
Publication Year :
2014
Publisher :
Natural Sciences Publishing, 2014.

Abstract

How to test the effect of density estimation methods is the key problem in the statistics. This paper presents a new criterion for assessing the effect of density estimation to select the suitable density e stimation method, using the maximum-entropy non-Gaussian measurement. Comparing withχ 2 -test and Dn-test, the method avoids the problem of the data interval division, and it is su itable for any type probability distribution. Simulation results show that the proposed me thod can accurately discriminate the pros and cons of different density estimation methods..

Details

ISSN :
23250399 and 19350090
Volume :
8
Database :
OpenAIRE
Journal :
Applied Mathematics & Information Sciences
Accession number :
edsair.doi...........3882d4aa760fb27be5a5e00c7b81be7b