103學年度第二學期專題討論統計學術演講公告
統計碩士學位學程將於104年3月31日(星期二)舉辦統計學術演講，邀請到中興大學統計學研究所陳律閎助理教授蒞臨演講，
講題：Principal component analysis for d-dimensional functional/longitudinal data
講員：陳律閎 助理教授 中興大學統計學研究所
時間：2015年3月31日星期二 下午 3：30 ~ 5：30
地點：臺灣大學博雅教學館409教室
費用：免費
名額：24位
摘要：
Functional principal component analysis (FPCA) serves as a fundamental toolbox for analyzing functional and longitudinal data. Various methods have been proposed for FPCA, however, most of the methods in the literature are designed for curves, i.e., functions with single variable.Recently, image data analysis draws lots of attention in the Statistics society. Image data can be treated as multidimensional functional data since they are usually smooth (or at least piecewisely smooth) and of infinite dimension. Thus, extending FPCA for multidimensional functional data, i.e., functions with multiple variables, such as images becomes an emerging issue.
In this work we extend the local polynomial based FPCA by Yao et al. (2005) to general d-dimensional functional and longitudinal data. Asymptotic results are presented, and our finite example experiments suggest that our approach significantly outperforms existing dimension reduction techniques for image data such as multilinear PCA.
歡迎對於此議題有興趣的師生同仁參加。
聯絡人：張仲凱
聯絡方式：E-mail：ntustat@ntu.edu.tw
TEL：(02) 3366-1481 FAX：(02) 3366-1482