### FANHUI KONG 1 AND HELIANG FEI 2

1 Department of Mathematics and Statistics, University of Maryland Baltimore County,
Baltimore, MD 21228, U.S.A.

2 Department of Mathematics, Shanghai Normal University, Shanghai, 200234, China

(Received February 7, 1995; revised August 7, 1995)

Abstract.    Assume n items are put on a life-time test, however for various reasons we have only observed the r1-th, ....., rk-th failure times xr1,n, ....., xrk,n with 0 < xr1,n < ··· < xrk,n < \infty. This is a multiply Type II censored sample. A special case where each xri,n goes to a particular percentile of the population has been studied by various authors. But for the general situation where the number of gaps as well as the number of unobserved values in some gaps goes to \infty, the asymptotic properties of MLE are still not clear. In this paper, we derive the conditions under which the maximum likelihood estimate of theta is consistent, asymptotically normal and efficient. As examples, we show that Weibull distribution, Gamma and Logistic distributions all satisfy these conditions.

Key words and phrases:    Maximum likelihood estimation, multiply Type II censoring, law of large numbers, central limit theorem, order statistic.

Source ( TeX , DVI , PS )