報(bào)告人簡介
鄒長亮，南開大學(xué)統(tǒng)計(jì)與數(shù)據(jù)科學(xué)學(xué)院教授、統(tǒng)計(jì)研究院院長。主要從事統(tǒng)計(jì)學(xué)及其與數(shù)據(jù)科學(xué)領(lǐng)域的交叉研究和實(shí)際應(yīng)用。研究興趣包括：高維數(shù)據(jù)統(tǒng)計(jì)推斷、變點(diǎn)和異常點(diǎn)檢測、預(yù)測性推斷等，在統(tǒng)計(jì)學(xué)和機(jī)器學(xué)習(xí)領(lǐng)域的權(quán)威雜志/會(huì)議Ann.Stat.、Biometrika、J.Am.Stat.Asso.、J.Mach.Learn.Res.、Math.Program.、ICML/NIPS/AAAI等上發(fā)表論文數(shù)十篇篇，入選愛思唯爾“中國高被引學(xué)者”。主持重大項(xiàng)目課題和科技部重點(diǎn)研發(fā)計(jì)劃課題等。任教育部科技委委員、全國應(yīng)用統(tǒng)計(jì)專業(yè)碩士教學(xué)指導(dǎo)委員會(huì)委員、中國現(xiàn)場統(tǒng)計(jì)研究會(huì)副理事長等。
內(nèi)容簡介
In multiple changepoint analysis, assessing the uncertainty of detected changepoints is crucial for enhancing detection reliability—a topic that has garnered significant attention. Despite advancements through selective p-values, current methodologies often rely on stringent assumptions tied to specific models and algorithms, potentially compromising the accuracy of post-detection statistical inference. We introduce TUNE (Thresholding Universally and Nullifying change Effect), a novel algorithm-agnostic approach that uniformly controls error probabilities across detected changepoints. TUNE sets a universal threshold for multiple test statistics, applicable across a wide range of algorithms, and directly controls the family-wise error rate without the need for selective p-values. Through extensive theoretical and numerical analyses, TUNE demonstrates versatility, robustness, and competitively power, offering a viable and reliable alternative for model-agnostic post-detection inference.
