課程大綱

課程資訊

課程名稱	資料科學之統計基礎(二) Statistical Foundations of Data Science
開課學期	106-2
授課對象	理學院應用數學科學研究所
授課教師	陳素雲
課號	MATH5079
課程識別碼	221 U8260
班次
學分	3.0
全/半年	半年
必/選修	選修
上課時間	星期三6,7,8(13:20~16:20)
上課地點	天數201
備註	資料科學學程課程。限本系所學生(含輔系、雙修生) 或限電資學院學生(含輔系、雙修生) 總人數上限：30人
Ceiba 課程網頁	http://ceiba.ntu.edu.tw/1062MATH5079_SFDS2
課程簡介影片
核心能力關聯	本課程尚未建立核心能力關連
課程大綱
為確保您我的權利,請尊重智慧財產權及不得非法影印
課程概述	This course will cover the following topics: unsupervised dimension reduction (SVD, PCA), supervised dimension reduction (sliced inverse regression SIR), tensor methods (high order SVD, multilinear PCA, multilinear algebra, tensor SIR, tensor regression), clustering analysis (k-means, self-updating process), discriminant analysis (LDA, logistic regression), kernel machines (kernel PCA, kernel SIR, kernel Fisher discriminant analysis, support vector machine, reproducing kernel Hilbert space), robust loss functions (Kullback-Leibler divergence, Bregman divergence, gamma-divergence, etc), neural networks (universal approximation theory, back-propagation, activation functions, dropout regularization).
課程目標	selected statistical methods and theory for machine learning and data science
課程要求	1. grading: 80% assignments (homework, mini projects), 20% class participation (attendance, in-class discussions) 2. able to write scripts for data analysis in a package such as Matlab, R, or Python 3. good understanding of theory is required
預期每週課後學習時數
Office Hours	另約時間備註： by appointment
指定閱讀	will be assigned in class
參考書目	An Introduction to Statistical Learning by James, Witten, Hastie and Tibshirani. web link to an ebook is posted on ceiba. The Elements of Statistical Learning: Data Mining, Inference, and Prediction by Hastie, Tibshirani and Friedman. web link to an ebook is posted on ceiba.
評量方式 (僅供參考)

課程進度

週次	日期	單元主題
第1週	2/28	national holiday, no class
第2週	3/07	1. course overview; 2. clustering analysis: k-means, hierarchical clustering and self-updating process for clustering ebooks: An Introduction to Statistical Learning http://www-bcf.usc.edu/~gareth/ISL/ The Elements of Statistical Learning
第3週	3/14	unsupervised dimension reduction: SVD, PCA, stability analysis
第4週	3/21	unsupervised dimension reduction (continued)
第5週	3/28	high order SVD
第6週	4/04	national holiday, no class
第7週	4/11	Tensor Toolbox. Tensor methods.
第8週	4/18	tensor methods; supervised dimension reduction, sliced inverse regression, KSIR Toolbox
第9週	4/25	linear discriminant analysis, logistic regression, maximum margin linear classifier
第10週	5/02	數學系自主學習週。複習 HOSVD, MPCA, LDA, logistic reg, linear SVM, 以及介紹軟體
第11週	5/09	robust loss functions and various divergence measures
第12週	5/16	kernel machines: kernel trick, kernel PCA, kernel Fisher discriminant analysis
第13週	5/23	kernel machines: SVM
第14週	5/30	1. kernel machines (continued) 2. Bring a laptop to class.
第15週	6/06	introduction to neural networks, universal approximation theory, back-propagation algorithm
第16週	6/13	neural networks (continued)
第17週	6/20	mini-project presentation (each person 8-10 minutes) 報告次序：黃聖堯，曾華廷，邱郁軒，劉妍君，鄭宗哲，森元俊成，朱傑韜，盧俊澎，林致弘，何文劭，洪嘉鴻，呂融昇，林伯儒，謝君宥，顏惠萱上 ceiba 確認你的作業 (1,2,3,6,7,8次的作業) 成績