Robust Estimates for Parameter of Exponential Distribution on Subrange

Abstract

The popular classical estimates for parameters of distributions are highly influenced by outliers, to avoid the effect of these outliers; robust estimates for parameters should be located gives approximate value of the classical estimation when the data are free from outliers.In this paper, suggestion was made to use (〖〖MAD〗_∝^*〗_ /D,D>0) as a robust estimates for the parameter of the exponential distribution. A simulation technique was used to generate data in different size based on random number generation to compare between the suggested robust estimates for different value of (∝) with the standard deviation depending on the mean square error (MSE). Results indicate that (〖〖MAD〗_0.4^*〗_ /0.327) is the best robust estimator for the parameter of exponential distribution of different size for the samples generated and for different value of the parameter.