Robust estimation in truncated discrete distributions with application to capture-recapture experiments (Q1101158)

A family of estimators is proposed for the relative frequency of the unobservable zero class in a truncated Poisson distribution. Large sample behavior is described by asymptotic expansion and robust properties are is a subvector of X with dimension \(p_ j\) such that \(\sum^{r}_{j=1}p_ j=p_ 0.\) Some admissible tests are derived for testing \(H_ 0:\mu =0\) versus \(H_ 1:\mu\neq 0\) based on a sample drawn from the whole vector X of dimension p and r additional samples drawn from \(X_{(1)},X_{(2)},...,X_{(r)}\), respectively. All \((r+1)\) samples are assumed to be independent. The distribution of some of the tests' statistics involved are also derived.