Nonparametric estimation of hazard rate under the constraint of monotonicity

摘要: This article shows how to smoothly "monotonize" standard kernel estimators of hazard rate, using bootstrap weights. Our method takes a variety forms, depending on choice estimator and the distance function used define certain constrained optimization problem. We confine attention particularly simple approach explore range functions. It is straightforward reduce "quadratic" inequality constraints "linear" equality constraints, so our may be implemented little more than conventional Newton-Raphson iteration. Thus, necessary computational techniques are very familiar statisticians. show both numerically theoretically that monotonicity, in either direction, can generally imposed rate regardless monotonicity or otherwise true rate. The case censored data easily accommodated. methods have extension problem testing for where plays role test statistic.