嘉宾：王德伟助理教授(南卡大学统计系)

时间：2015年12月21日下午4:00-5:00

地点：管理科研大楼1008会议室

题目：Nonparametric goodness-of-fit tests for uniform stochastic ordering

摘要：We propose an L p distance-based family of goodness-of-fit (GOF) tests for uniform stochastic ordering with two continuous distributions F and G, both of which are unknown. Our tests are motivated by the fact that when F and G are uniformly stochastically ordered, the ordinal dominance curve (ODC) R = F G−1 is star-shaped. Therefore, the L p distance between the sample ODC and its least star-shaped majorant should be small when F and G satisfy uniform stochastic ordering and large otherwise. We derive asymptotic distributions and prove that our testing procedure has a unique least favorable configuration of F and G for all p ∈ [1,∞]. We use simulation to assess finite-sample properties and demonstrate that a modified, one-sample version of our procedure (e.g., with G known) is more powerful than the one-sample GOF test suggested by Arcones and Samaniego (2000). We also discuss sample size determination. We illustrate our methods using data from a pharmacology study evaluating the effects of administering caffeine to prematurely born infants.

 

欢迎感兴趣师生参加！