Classiication for High-dimension Small-sample Data Sets Based on Kullback-leibler Information Measure
نویسندگان
چکیده
In classifying samples by Gaussian clas-siier, the covariance matrix estimated with a small number sample set becomes unstable, which leads to degrading the classiication accuracy. In this paper , we discuss the covariance matrix estimation problem for small number samples with high dimension setting based on Kullback-Leibler Information Measure. A new covariance matrix estimator is developed, and a fast, rough estimating regular-ization parameter formula is derived. Experiments are performed to investigate the classiication accuracy with developed covariance matrix estimator and higher classiication accuracy results are obtained .
منابع مشابه
Using Kullback-Leibler distance for performance evaluation of search designs
This paper considers the search problem, introduced by Srivastava cite{Sr}. This is a model discrimination problem. In the context of search linear models, discrimination ability of search designs has been studied by several researchers. Some criteria have been developed to measure this capability, however, they are restricted in a sense of being able to work for searching only one possibl...
متن کاملMulti-regularization Parameters Estimation for Gaussian Mixture Classifier based on MDL Principle
Regularization is a solution to solve the problem of unstable estimation of covariance matrix with a small sample set in Gaussian classifier. And multi-regularization parameters estimation is more difficult than single parameter estimation. In this paper, KLIM_L covariance matrix estimation is derived theoretically based on MDL (minimum description length) principle for the small sample problem...
متن کاملModel Confidence Set Based on Kullback-Leibler Divergence Distance
Consider the problem of estimating true density, h(.) based upon a random sample X1,…, Xn. In general, h(.)is approximated using an appropriate in some sense, see below) model fƟ(x). This article using Vuong's (1989) test along with a collection of k(> 2) non-nested models constructs a set of appropriate models, say model confidence set, for unknown model h(.).Application of such confide...
متن کاملDiscriminant Analysis for ARMA Models Based on Divergency Criterion: A Frequency Domain Approach
The extension of classical analysis to time series data is the basic problem faced in many fields, such as engineering, economic and medicine. The main objective of discriminant time series analysis is to examine how far it is possible to distinguish between various groups. There are two situations to be considered in the linear time series models. Firstly when the main discriminatory informati...
متن کاملTesting Exponentiality Based on Renyi Entropy of Transformed Data
In this paper, we introduce new tests for exponentiality based on estimators of Renyi entropy of a continuous random variable. We first consider two transformations of the observations which turn the test of exponentiality into one of uniformity and use a corresponding test based on Renyi entropy. Critical values of the test statistics are computed by Monte Carlo simulations. Then, we compare p...
متن کامل