同济大学数学系学术报告

题 目：Some new research based on sequential methods

报告人：濮晓龙教授

(华东师范大学金融与统计学院）

【Abstract】For sequential test plans, we propose the weighted expected sample size (WESS) to evaluate the overall performance of sequential test plans on the parameters of interest. Motivated by minimizing the WESS to control the expected sample sizes, we develop three sequential mixture likelihood ratio tests for one-sided composite hypotheses. It is proved that the proposed tests are the asymptotically optimal and have a finite stopping time under some conditions. Simulation results provide the evidence that the proposed tests have some superiority over the SPRT and 2-SPRT.

Sensitivity tests are used to make inferences about a sensitivity, a characteristic property of some products that cannot be observed directly. Aiming at improving both the speed of getting an overlap and the estimation precision, we propose two sequential sensitivity tests. Our simulation studies indicate that the proposed tests outperforms the Langlie, Neyer and Dror-Steinberg tests. Moreover, the superiority of the proposed tests are confirmed by two real applications.

Based on the weighted-loss-function, a new single EWMA chart is proposed to detect both mean and variance shifts simultaneously. It includes the EWMA control chart based on the SC statistic and Weighted-Loss-Function control chart as its special cases. The performance of the proposed method is illustrated by simulations and a real data example.

时间：2013年12月19日（周四）下午16:00开始

地点：数学系致远楼107室

欢迎广大师生前来参加