報告人: 蔣學(xué)軍副教授 南方科技大學(xué)
時間：2023年11月17日 10:30-12:00
地點(diǎn)：數(shù)學(xué)與信息學(xué)院201報告廳
Abstract: Strong correlation among predictors poses a great challenge in the analysis of ultra-high dimensional data. This leads to an increase in the computation time for screening active variables and a decrease in selection accuracy. To address this issue, we propose an innovative two-stage feature selection procedure and its derivative versions based on sparsity assumption. This approach initially screens important features by ranking quantile ridge estimation and subsequently employs a likelihood-based post-screening selection strategy to refine variable selection. This procedure is simple to implement and maintains the robustness of quantile regression. In addition, it possesses other numerous other desirable properties from theoretical and computational standpoints. Theoretically, we establish the strong consistency of feature selection for the proposed methods under some regularity conditions. In empirical studies, we assess the finite sample performance of our methods by comparing them with marginal quantile utility screening approaches and existing penalized quantile regression methods. Furthermore, we apply our methods to identify genes associated with anticancer drug sensitivities.
報告人簡介: 蔣學(xué)軍，南方科技大學(xué)統(tǒng)計(jì)與數(shù)據(jù)科學(xué)系副教授（長聘），博士生導(dǎo)師，于2009年博士畢業(yè)于香港中文大學(xué)統(tǒng)計(jì)系，2009-2010在港中文從事博士后研究，2010-2013任中南財(cái)經(jīng)政法大學(xué)副教授，于2013年07月加入南方科技大學(xué)，入選深圳市海外高層次人才孔雀計(jì)劃（2016），曾獲南方科技大學(xué)杰出教學(xué)獎（2018），深圳市優(yōu)秀教師（2018），主持和完成國家（廣東省）自然科學(xué)基金、深圳市基礎(chǔ)研究面上項(xiàng)目等10余項(xiàng)。其主要研究方向包括分位數(shù)回歸、變量選擇、假設(shè)檢驗(yàn)、高維統(tǒng)計(jì)推斷，金融統(tǒng)計(jì)與計(jì)量等，已發(fā)表包括Bernoulli Journal , Statistica Sinica, Econometrics Journal, Science China-Mathematics等在內(nèi)的SCI&SSCI論文50余篇，授權(quán)專利1項(xiàng)，并出版英文教材一部。國內(nèi)學(xué)會任職主要有中國現(xiàn)場統(tǒng)計(jì)研究會-教育統(tǒng)計(jì)與管理分會副理事長，多元分析應(yīng)用專業(yè)委員會秘書長等。

歡迎廣大師生參加！