International Journal of Nonlinear Analysis and Applications

A comprehensive assessment of the performance and stability of maximum likelihood estimators for the exponential–poisson distribution: An analysis of sample size and censoring schemes on the NR, EM, and SEM algorithms under type I progressive interval censoring

Articles in Press

Document Type : Research Paper

Authors

Department of Statistics, Payame Noor University (PNU), P.O.Box 19395-4697, Tehran, Iran

Abstract

This study compares Newton-Raphson (NR), EM, and stochastic EM (SEM) for estimating Exponential-Poisson parameters under Type I progressive interval censoring. Simulations across sample sizes (20–200) and censoring schemes show SEM achieves the smallest MSE and highest stability, especially for small samples or complex censoring. EM performs reliably for large samples but is more variable for small samples. NR is highly sensitive to initial values and least stable. A real melanoma dataset confirms EP distribution fit (KS=0.0961) and replicates simulation patterns, emphasizing the need to choose estimation methods carefully.

Keywords

[1] K. Adamidis and S. Loukas, A lifetime distribution with decreasing failure rate, Stat. Probab. Lett. 39 (1998), 35-42.
[2] R. Aggarwala, Progressive interval censoring: some mathematical results with applications to inference, Commun. Stat. Theory Meth. 30 (2001), 1921-1935.
[3] R. Arabi Belaghi, M. Noori Asl, and S. Singh, On estimating the parameters of the Burr XII model under progressive type-I interval censoring, J. Stat. Comput. Simul. 87 (2017), no. 16, 3132-3151.
[4] N. Balakrishnan and R. Aggarwala, Progressive Censoring: Theory, Methods, and Applications, Springer Science & Business Media, 2000.
[5] S. Carbon, A. Ireland, C.J. Mungall, S. Shu, B. Marshall, S. Lewis, AmiGO Hub, and Web Presence Working Group, AmiGO: online access to ontology and annotation data, Bioinformatics 25 (2009), no. 2, 288-289.
[6] D.G. Chen and Y.L. Lio, Parameter estimations for generalized exponential distribution under progressive type-I interval censoring, Comput. Stat. Data Anal. 54 (2010), 1581-1591.
[7] S. Dey, S. Singh, and Y.M. Tripathi, Estimation and prediction for a progressively censored generalized inverted exponential distribution, Stat. Methodol. 32 (2016), 185-202.
[8] J. Lawless, Statistical Models and Methods for Lifetime Data, New York: John Wiley & Sons, 1982.
[9] C.T. Lin, S.J.S. Wu, and N. Balakrishnan, Planning life tests with progressively type-I interval censored data from the lognormal distribution, J. Stat. Plann. Inf. 139 (2009), 54-61.
[10] H.K.T. Ng and Z. Wang, Statistical estimation for the parameters of the Weibull distribution based on progressively type-I interval censored sample, J. Stat. Comput. Simul. 79 (2009), 145-159.
[11] M. Noori Asl, R.A. Arabi Belaghi, and H. Bevrani, On Burr XII distribution analysis under progressive type-II hybrid censored data, Methodol. Comput. Appl. Probab. 19 (2017), no. 2, 665-683.
[12] M. Noori Asl, R. Arabi Belaghi, and H. Bevrani, Classical and Bayesian inferential approaches using Lomax model under progressively type-I hybrid censoring, J. Comput. Appl. Math. 343 (2018), 397-412.
[13] B. Pradhan and E.V. Gijo, Parameter estimation of lognormal distribution under progressive type-I interval censoring, Technical report no. sqcor-2013-02, 2013.
International Journal of Nonlinear Analysis and Applications

Articles in Press, Corrected Proof
Available Online from 06 May 2026
  • Receive Date: 03 December 2025
  • Revise Date: 02 February 2026
  • Accept Date: 02 February 2026