A new entropy estimator and its application to goodness of fit test for Weibull distribution

Abstract

In this article, we introduce a new estimator of entropy of continuous random variables. Bias, variance and the mean squared error of the new estimator are obtained and compared with the other existing estimators. The results show that the proposed estimator has a lower mean squared error than its competitors. Then, we propose some goodness of fit tests for Weibull distribution based on the entropy estimators. To assess the effectiveness of the proposed tests, we utilize Monte Carlo simulation to evaluate their power against eighteen different alternatives with varying sample sizes. The results show that the tests are powerful and we can use them in practice. Finally, two real datasets are considered and modeled by the Weibull distribution.