報(bào) 告 人：崔恒建 教授
報(bào)告題目：Model-free feature screening based on Hellinger distance for ultrahigh dimensional data
報(bào)告時(shí)間：2024年3月29日（周五）下午2：30
報(bào)告地點(diǎn)：騰訊會(huì)議 243-343-042
主辦單位：數(shù)學(xué)研究院、數(shù)學(xué)與統(tǒng)計(jì)學(xué)院、科學(xué)技術(shù)研究院
報(bào)告人簡(jiǎn)介：
崔恒建，現(xiàn)為首都師范大學(xué)教授，博士生導(dǎo)師，中國(guó)科協(xié)第十屆全委會(huì)委員，曾任國(guó)務(wù)院學(xué)位委員會(huì)學(xué)科評(píng)議組專(zhuān)家。中國(guó)科學(xué)院系統(tǒng)科學(xué)研究所博士畢業(yè)。在大數(shù)據(jù)統(tǒng)計(jì)建模、高維統(tǒng)計(jì)及其穩(wěn)健統(tǒng)計(jì)理論和方法、統(tǒng)計(jì)機(jī)器學(xué)習(xí)、金融統(tǒng)計(jì)、以及質(zhì)量管理等領(lǐng)域取得過(guò)許多重要的研究成果，發(fā)表論文180余篇，其中包括發(fā)表在國(guó)際頂級(jí)的統(tǒng)計(jì)和計(jì)量經(jīng)濟(jì)學(xué)雜志JASA、AoS、JRSS(B)、Biometrika和JoE上。主持國(guó)家自然科學(xué)基金重點(diǎn)項(xiàng)目、杰青（B）項(xiàng)目以及多項(xiàng)面上項(xiàng)目、主要參加教育部重大科研基金項(xiàng)目、科技部863等項(xiàng)目?，F(xiàn)擔(dān)任《數(shù)學(xué)學(xué)報(bào)》和《應(yīng)用數(shù)學(xué)學(xué)報(bào)》中、英文版以及《Statistical Theory and Related Fields》編委，中國(guó)現(xiàn)場(chǎng)統(tǒng)計(jì)研究會(huì)副理事長(zhǎng)，全國(guó)工業(yè)統(tǒng)計(jì)教育研究會(huì)副理事長(zhǎng)，北京應(yīng)用統(tǒng)計(jì)學(xué)會(huì)會(huì)長(zhǎng)，國(guó)際數(shù)理統(tǒng)計(jì)學(xué)會(huì)（中國(guó)分會(huì)）常務(wù)理事。曾獲得教育部高等學(xué)?？茖W(xué)技術(shù)獎(jiǎng)-自然科學(xué)獎(jiǎng)二等獎(jiǎng)；全國(guó)統(tǒng)計(jì)科學(xué)研究?jī)?yōu)秀成果獎(jiǎng)一等獎(jiǎng)等。
報(bào)告摘要：
With the explosive development of data acquisition and processing technology, the dimension of features increases exponentially with the sample size, which poses great challenges for data analysis. It is vital to accurately identify useful features from thousands of them. In this paper, we develop an omnibus model-free feature screening procedure based on the Hellinger distance with some appealing merits. First, we define the Hellinger distance index for discrete response variables in discriminant analysis. Second, this procedure works consistently for continuous response variables, in which the continuous responses are discretized by slice-and-fused technique. Third, it is robust to the potential outliers and model misspecification. Theoretically, the procedure for discrete and continuous response variables possess sure screening properties and ranking consistency properties under mild conditions. Numerical studies demonstrate that this procedure exhibits strong competitiveness in heavy-tailed and skewed data, while remaining comparable to existing approaches for light-tailed data, indicating its robustness performance across a range of data. Real data contains two examples, discrete and continuous response variables, to illustrate the effectiveness of the proposed method.