报告题目：Linear Discriminant Analysis with High-dimensional Mixed Variables
报告地点：腾讯会议813 491 974
报告摘要：With the rapid development of modern measurement technologies, datasets containing both discrete and continuous variables are more and more commonly seen in different areas, and in particular, the dimensions of the discrete and continuous variables can oftentimes be very high. Though discriminant analysis for mixed variables under the traditional fixed dimension setting has been well studied since the 80's, promising approaches taking into account both the high dimensionality and the mixing nature of the data sets are still missing. In this paper, we aim to develop a simple yet useful classification rule that addresses both the high dimensionality and the mixing structure of the variables simultaneously. Our framework is built on a location model, under which we further propose a semiparametric formulation for the optimal Bayes rule. We show that the optimal classification direction and the intercept in the optimal rule can be estimated separately. Efficient direct estimation schemes are then developed to obtain consistent estimators of the discriminant components. Asymptotic results on the estimation accuracy and the misclassification rates are established, and the competitive performance of the proposed classier is illustrated by simulation and real data studies.
主讲人简介：蒋滨雁教授，2007年本科毕业于中国科学技术大学，2012年于新加坡国立大学获得统计学博士学位，随后在美国卡内基梅隆大学做博士后研究学者。现为香港理工大学应用数学系副教授。其丰硕的研究成果多发表在Biometrika, Journal of the American Statistical Association, Computational Statistics and Data Analysis, Statistica Sinica, Journal of Machine Learning Research, Annals of the Institute of Statistical Mathematics, Electronic Journal of Statistics等统计学和机器学习等顶级或高质量期刊上。主要研究方向是高维数据分析，网络数据分析等。