内容标题36

  • <tr id='whD1ZD'><strong id='whD1ZD'></strong><small id='whD1ZD'></small><button id='whD1ZD'></button><li id='whD1ZD'><noscript id='whD1ZD'><big id='whD1ZD'></big><dt id='whD1ZD'></dt></noscript></li></tr><ol id='whD1ZD'><option id='whD1ZD'><table id='whD1ZD'><blockquote id='whD1ZD'><tbody id='whD1ZD'></tbody></blockquote></table></option></ol><u id='whD1ZD'></u><kbd id='whD1ZD'><kbd id='whD1ZD'></kbd></kbd>

    <code id='whD1ZD'><strong id='whD1ZD'></strong></code>

    <fieldset id='whD1ZD'></fieldset>
          <span id='whD1ZD'></span>

              <ins id='whD1ZD'></ins>
              <acronym id='whD1ZD'><em id='whD1ZD'></em><td id='whD1ZD'><div id='whD1ZD'></div></td></acronym><address id='whD1ZD'><big id='whD1ZD'><big id='whD1ZD'></big><legend id='whD1ZD'></legend></big></address>

              <i id='whD1ZD'><div id='whD1ZD'><ins id='whD1ZD'></ins></div></i>
              <i id='whD1ZD'></i>
            1. <dl id='whD1ZD'></dl>
              1. <blockquote id='whD1ZD'><q id='whD1ZD'><noscript id='whD1ZD'></noscript><dt id='whD1ZD'></dt></q></blockquote><noframes id='whD1ZD'><i id='whD1ZD'></i>

                西∞南财经大学主页

                当前位置:首页 > 西南财经」大学 > 新闻公告 >

                香港浸会大学 彭衡副教授:BOLT-SSI: Fully Screening Interaction Effects for Ultra-High Dimensional Data

                2019-11-06 0 新闻公告 来源:西南财¤经大学新闻网

                光华讲坛——社会名流与企业家论坛第 5585 期

                 

                主题:BOLT-SSI: Fully Screening Interaction Effects for Ultra-High Dimensional Data

                主讲人:香港浸会大学 彭衡副教授

                主持人:统】计学院统计研究中々心 林华珍教授

                时间:2019年11月07日(星期四)上午11:00-12:00

                地点:西南财经大学柳林校区弘远楼408会议室

                主办单位:统计研究中心 统计学院 科研处

                 

                主讲人简介:

                彭衡,现为香港浸会大学数学系副教授,2003年从香港中文大学取得█统计学博士学位,2003年-2006年在普林斯顿大学做博士后。他主要从事非参数与半◣参数模型、模型选择、高维数据建模、混合模型等领域的研究。他是IMS的会员,2011-2014担任Statistica Sinica副主编,现为Computational Statistics and Data Analysis副主编;曾做过Annals,JASA,JRSSB, Biometrika, Statistica Sinica等的评审。在统计学ㄨ国际顶级期刊Annals,JASA, Statistica Sinica,TEST,Computational Statistics and Data Analysis上发表论文十余篇。

                主要内容:

                Detecting interaction effects among predict variables to response variables is often an crucial step in regression modeling of real data for various applications. In this paper by marginal likelihood functions, we firstly introduce a simple sure screening procedure (SSI) to fully detect significant pure interaction between predict variables and the response variable in the high or ultra-high dimensional generalized linear regression models. Furthermore, we suggest to discretize continuous predict variables, and utilize the Boolean operation for the marginal likelihood estimates. The so called BOLT-SSI procedure is proposed to accelerate the sure screening speed of the procedure.  We investigate the sure screening properties of SSI and BOLT-SSI.   Our studies have several important features. First, to most ultra-high dimensional data in practice, the proposed sure screening methods can fully detect any pure interaction effects among ultra-high dimensional data. It is an impossible finished task from theoretical insight. Second, the proposed method efficiently takes the advantages of computer architecture to speed up the proposed algorithm and make trade-off between the computation burden and statistical modeling efficiency. Specially, regarding the interaction effect detecting study as a special example, by this study we show the limitation of theoretical investigation from the practical insight, and illustrate that how to make trade-off between engineering techniques and theoretical investigations.

                在大量实际应用中,检测预测变量与响应变量之间的←交互作用通常是回归建模中至关重要的一步。针对高维或〇超高维广义线性模型,本文基于边际似然函数,首先介绍一种简单的确定筛选算法(SSI),以完全检测预测变量和响应变量→之间显著的纯交互作用。此外,我们建议离散化连续型预测变量,并将布尔运算用于边际似然估计。本文提¤出了BOLT-SSI算法以加快确定筛选的速度,同时考查了SSI和BOLT-SSI的确定筛选性质。我们的研究有几个重要特征。首先,对于实际中的大多♀数超高维数据,所提出的确◎定筛选方法可以完全地检测超高维数据之间的任何纯交互作用,从理论上◥看这是不可能完成的任务。其次,所提出的方法有效地利用了计算机体系结构的『优势来加快算法的运算速度,并在庞大的计算和统计建模功效之间进行平衡。特别地,以交互∮作用检测研究为例,本报告从实践的角度揭示理论研究所具有的局限性,并说明如何在工程技术和理论研究之间进行权ㄨ衡。

                 


                未经允许不得转〓载:二九年华大学门户 » 香港浸会大学 彭衡副教授:BOLT-SSI: Fully Screening Interaction Effects for Ultra-High Dimensional Data

                相关推荐

                标签