報 告 人：張久軍 教授

報告題目：Distribution-free joint change-point detection of location and scale – a comprehensive reviewand some new results

報告時間：2025年4月19日（周六）下午2點

報告地點：靜遠樓1506學術(shù)報告廳

主辦單位：數(shù)學與統(tǒng)計學院、數(shù)學研究院、科學技術(shù)研究院

報告人簡介：

       張久軍，遼寧大學數(shù)學與統(tǒng)計學院教授，博士研究生導師，統(tǒng)計與數(shù)據(jù)科學系主任。遼寧大學“雙帶頭人”教師黨支部書記，美國明尼蘇達大學訪問者，遼寧省“百千萬人才工程”百人層次人選，遼寧省普通本科教學名師，遼寧大學本科優(yōu)秀主講教師。兼任中國現(xiàn)場統(tǒng)計研究會多元統(tǒng)計分析分會理事，中國現(xiàn)場統(tǒng)計學會生存分析分會理事，遼寧省數(shù)學會理事。近年來，主持、參與國家自然科學基金，遼寧省自然科學基金十余項，發(fā)表論文50余篇。

報告摘要：

        Traditional process monitoring schemes often assume knowledge of process distributions. However, in the real world, where distribution information is rarely available, this necessitates the use of distribution-free schemes. Lepage and Cucconi statistics are commonly employed to construct nonparametric control schemes for jointly monitoring location and scale parameters. The change-point design (CPD) approach is frequently utilized for constructing monitoring schemes, as it does not require as many in-control (IC) reference sample observations compared to traditional Phase-II schemes. This paper investigates two new joint location-scale CPD schemes by integrating other Lepage-type statistics into the exponentially weighted moving average (EWMA) scheme. The study also compares these new schemes with existing Lepage and Cucconi-type CPD schemes. Simulation results are promising and support the application of these schemes in industrial contexts. Finally, real-world data is used to illustrate the effectiveness of the various schemes.