A weighted method for the exclusive hypothesis test with application to typhoon data

• DOI: 10.1002/cjs.11618
• Publication Date: 2021-05-22T08:40:28Z
• Authors: Yi Wang; Peng Wu; Xingwei Tong; Jianguo Sun
• Keywords: Alternative hypothesis; Bayesian information criterion; False Discovery Rate; Multiple comparisons problem; Null (SQL); Pleiotropy; Divergence (linguistics)

Motivated by the testing of genetic pleiotropy, we discuss a general class hypothesis testing, exclusive test (EHT). A is an EHT if null can be divided into set sub‐hypotheses, and main difficulty for calculation p ‐value. To address this problem, propose weighted procedure develop two methods, one likelihood‐based other Bayesian information criterion (BIC)‐based, determining corresponding weights. Furthermore, show that BIC‐based method control asymptotic type I error. We conduct extensive simulation study these proposed which suggests they work well in practice. In particular, new shown both theoretically numerically to exhibit better performance than existing two‐stage decision rule pleiotropy. Our methodology then applied data concerning tropical storms.