Department of Statistics, The Ohio State University
Statistics and Biostatistics Colloquium Series

Since its introduction in the sixties progressive censoring has received great attention in the literature. In particular, many results have been established in the last decade. Progressively censored order statistics from a sample $X_1, X_2, \ldots , X_n$ can be constructed as follows: Suppose that $X_1, X_2,\ldots , X_n$ are random lifetimes corresponding to $n$ independent units in a life-testing experiment. These random variables are observed in increasing order. Now, at times $T_1,\ldots, T_m$, pre-determined numbers $R_1,\ldots, R_m$ of surviving items are randomly withdrawn from the sample. In standard progressive censoring, the times $T_i$ are fixed (so-called Type-I censoring) or defined as the minimum of the life-times of those units which have not been failed or progressively censored from the sample (so-called Type-II censoring). In this talk, we review some basic results in the standard models of progressive censoring. Then, some extensions regarding both the censoring procedure (i.e., hybrid progressive censoring) and the distributional assumption (e.g., INID models) are presented. Finally, so-called adaptive progressive censoring is introduced. In this model, the censoring numbers are not fixed in advance and can depend on previous censoring times.

Meet the speaker in Room 212 Cockins Hall at 4:30 p.m. Refreshments will be served.