Sample entropy applied to test of fit for Lindley distribution and power comparison

Abstract

This work is devoted to construction of goodness-of-fit test statistics for the Lindley model. The proposed test statistics are based on the sample entropy. The properties of the tests and critical points of them for various sample sizes are obtained. An empirical power comparison study is then conducted using Monte Carlo simulations, under various alternatives and for different sample sizes. The new tests exhibit good performance in terms of power in comparison to the EDF-based test statistics. A discussion of the results follows, where the best tests are selected for different classes of alternatives. Finally, a real dataset is analyzed for illustrative purposes.