Non-strange weird resampling for complex survival data

摘要

This paper introduces a new data-dependent multiplier bootstrap for nonparametric analysis of survival data, possibly subject to competing risks. The new procedure includes the general wild bootstrap and the weird bootstrap as special cases. The data may be subject to independent right-censoring and left-truncation. The asymptotic correctness of the proposed resampling procedure is proven under standard assumptions. Simulation results on time-simultaneous inference suggest that the weird bootstrap performs better than the standard normal multiplier approach.