STOCHASTIC EXPECTATION MAXIMIZATION ALGORITHM FOR EXPONENTIAL-POISSON DISTRIBUTION UNDER TYPE-I PROGRESSIVE INTERVAL CENSORING

STOCHASTIC EXPECTATION MAXIMIZATION ALGORITHM FOR EXPONENTIAL-POISSON DISTRIBUTION UNDER TYPE-I PROGRESSIVE INTERVAL CENSORING

A. Mohammadi, A. Shadrokh, M. Yarmohammadi

[PDF]

Abstract

The Exponential-Poisson (EP) distribution is generated by combining the Exponential distribution with a zero truncated Poisson distribution as a model for lifetime data with decreasing failure rate. This paper deals with the problem of estimating unknown parameters of the Exponential-Poisson distribution as a lifetime model when samples are observed under progressive type-I interval censoring. We employ the Newton-Raphson (NR), classical expectation maximization (EM) and stochastic expectation maximization (SEM) algorithms to find the maximum likelihood estimates for the unknown parameters. The performance of the proposed SEM estimators are illustrated by a Monte Carlo simulation study and used for a real data set. Our simulation showed that the performance of SEM algorithm is quite satisfactory on the basis of mean square error and by increasing the sample size, the efficiency is also increases.

Keywords

Maximum likelihood estimation, Progressive Interval Censoring, EM and SEM algorithm, Exponential Poisson.