A simulation study of likelihood inference procedures in Rayleigh distribution with censored data

Abstract EN

Inference procedures based on the likelihood function are considered for the one parameter Rayleigh distribution with typel and type 2 censored data.

Using simulation techniques, the finite sample performances of the maximum likelihood estimator and the large sample likelihood interval estimation procedures based on the Wald, the Rao, and the likelihood ratio statistics are investigated.

It appears that the maximum likelihood estimator is unbiased.

The approximate variance estimates obtained from the asymptotic normal distribution of the maximum likelihood estimator are accurate under type 2 censored data, while they tend to be smaller than the actual variances when considering type 1 censored data of small size.

It appears also that interval estimation based on the Wald and Rao statistics need much more sample size than interval estimation based on the likelihood ratio statistic to attain reasonable accuracy.