报告名称：Change detection in over-dispersed network data using the Grouped Negative-Binomial Network Autoregressive Model

报告人：张久军 教授 辽宁大学

报告时间：2025年11月2日 下午1:30--2:30

报告地点：伍卓群楼第2报告厅

校内联系人：李聪 [email protected]

报告摘要：

Automotive complaint data typically encompass a wide range of subject areas, where various topics of defects are often interrelated rather than occurring in isolation. Such dependency structures can be naturally captured through network representations. In this context, modern quality monitoring in complex systems requires the analysis of over-dispersed integer-valued time series with network-based dependencies. Conventional methods often underperform, as they generally fail to account for network topology and node-level heterogeneity. To address these limitations, we propose a Grouped Negative-Binomial Network Autoregressive (G-NBNAR) model that incorporates group-specific parameters to capture heterogeneous nodal dynamics. The theoretical properties of the model—including stationarity and estimator consistency—are rigorously established. For practical implementation, the model is integrated with a top-r CUSUM monitoring scheme. Extensive simulations and an empirical study using real automobile complaint data demonstrate the superiority of the proposed framework over existing methods, offering a statistically sound tool for quality monitoring in networked environments.

个人简介：

张久军，辽宁大学数学与统计学院教授，博士研究生导师，统计与数据科学系主任。辽宁大学“双带头人”教师党支部书记，美国明尼苏达大学访问者。辽宁省第13批“百千万人才工程”百人层次人选，“沈阳市高级人才-领军人才”，辽宁省普通本科教学名师，辽宁省高校“校园先锋示范岗”（个人）。曾被评为辽宁大学本科优秀主讲教师、辽宁大学师德师风模范标兵等称号。兼任中国现场统计研究会多元统计分析分会理事，中国现场统计学会生存分析分会理事，辽宁省数学会理事。主持、参与国家自然科学基金，辽宁省自然科学基金十余项，发表论文50余篇。