Testing for Equality of Distributions Using Empirical Likelihood Approach

Event Start
Event End
Location
Building 2, Level 5, Room 5209
Dr. Zhi LIU, Assistant Professor in Department of Mathematic of University of Macau, China

Abstract

The comparison of multivariate populations has been involved in many fields. For example, in testing the efficiency of a new cure, we would compare it with the usual treatment. In this work, we investigate the problem of testing the equality of distribution of two populations. We combine the characteristic functions and empirical likelihood approach. It has been proved that the test statistic tends to a chi-square distribution. Simulation studies are presented to study the finite-sample performance of our method. A real data set is analyzed to illustrate our method.

Brief Biography

The comparison of multivariate populations has been involved in many fields. For example, in testing the efficiency of a new cure, we would compare it with the usual treatment. In this work, we investigate the problem of testing the equality of distribution of two populations. We combine the characteristic functions and empirical likelihood approach. It has been proved that the test statistic tends to a chi-square distribution. Simulation studies are presented to study the finite-sample performance of our method. A real data set is analyzed to illustrate our method.