Estimation and classification using progressive type-II censored samples from two exponential populations with a common location

This article considers the problems of estimation and classification using progressive type-II censored samples as training data from two exponential populations with a common location and different scale parameters. First, we derive improved estimators over the maximum likelihood estimator (MLE) and uniformly minimum variance unbiased estimator (UMVUE) of the common location parameter with and without order restriction on the scale parameters. Then using all those estimators, we construct several classification rules to classify a single or a group of observations into one of the two exponential populations. The probabilities of correct classification have been compared for all the rules. We perform a detailed simulation study to evaluate the performances of all the rules in terms of the expected probability of correct classification (EPC) numerically. Finally, a real-life example is considered for application purposes.