Power Comparison between Methods of Empirical Process and a Kernel Density Estimator for the Test of Distribution Change Na, Seong-Ryong; Park, Hyeon-Ah;
There are two nonparametric methods that use empirical distribution functions and probability density estimators for the test of the distribution change of data. In this paper we investigate the two methods precisely and summarize the results of previous research. We assume several probability models to make a simulation study of the change point analysis and to examine the finite sample behavior of the two methods. Empirical powers are compared to verify which is better for each model.
Test for distribution change;empirical distribution function;probability density estimator;nonparametric test;simulation study;