Estimators of a distribution function with increasing failure rate average (Q1107925): Difference between revisions

A nonparametric estimator \(G_ n\) of a distribution function F with increasing failure rate average (IFRA) is constructed. This estimator \(G_ n\) is so constructed that its hazard function \(C_ n\) is the greatest star-shaped minorant of the sample hazard function. The estimator \(G_ n\) is itself IFRA. Denoting the sample distribution function by \(F_ n\) and the sample hazard function by \(H_ n\), it follows from \textit{P. W. Millar} [Z. Wahrscheinlichkeitstheorie verw. Gebiete 48, 233-252 (1979; Zbl 0387.62029)] that \(F_ n\) is an asymptotically minimax estimator of F if F is IFRA. We prove, under suitable restrictions on F, and for any fixed \(\lambda\) with \(F(\lambda)<1\), that \[ n^{1/2}\sup_{x\leq \lambda}| G_ n(x)-F_ n(x)| \quad and\quad n^{1/2}\sup_{x\leq \lambda}| C_ n(x)-H_ n(x)| \] both tend to zero in probability. That is, \(G_ n\) and \(F_ n\) are asymptotically \(n^{1/2}\)-equivalent as estimators of the true distribution functions. Similar results can be obtained for a DFRA distribution.