CS 534: Machine Learning 

Spring 2011      

---

Instructor: Xiaoli Fern

Email:                xfern@eecs.oregonstate.edu 
Office:               Kelly 3073
Office Hours:     MWM 11-12 (before class) or by appointment
Class email list:   cs534-sp11@engr.orst.edu


Course information (policy, learning objectives etc)

Quick links: Announcements, textbook and materials, assignments, lectures

---

Announcements

• Welcome to CS534! Please check this page often for new updates
• Per request from students, I have posted a bunch of slides from previous years. The class will roughly follow these slides, but minor (sometimes major) modifications may occur and the slides will be updated accordingly.
  
• Homework 1 written part solution set is posted! 4/15/11
• Final Project Information is posted! 4/20/11
• Midterm solution is posted! 05/09/11

---

Textbook and materials

• Recommended Text: Pattern recognition and machine learning, by Chris Bishop. 1st edition. (CB)　

• Other references: 
• Machine learning, by Tom Mitchell (TM)
• Pattern Classification, by Duda, Hart and Stork. 2nd Edition (DHS)
• Useful materials:
• A brief review of basic probablity concepts Courtesy of Andrew Ng
• A brief review of linear algebra and vector calculus
• Matrix Cookbook Great reference for matrix related derivations. 
• Useful video lectures
• Linear Algebra video lectures of Professor Gilbert Strang from MIT opencourseware
• Multivariable Calculus
  

---

Assignments

• HW1 : Due 4/11/11 written part solution set ; data: twogaussian and iris-twoclass
• HW2 : Due 4/29/11 written part solution ; Naive Bayes data: 20newsgroup and Decision tree data zoo-train and zoo test
• Written HW3 : Due 5/16/11; solution; Implementation assignment 3 : Due 5/20/11. Data: Spect train; Spect test
• Written HW4 : Due Friday June 3rd. solution

---

Lecture Schedule (note the materials posted below frequently updated, please note the time stamp)

• Week 1(3/28/11-4/1/11)
• Basic introduction (slides slightly updated 3/30/11)
• Review of probability and Maximum Likelihood Estimation (slides )
• Linear regression ( slides )

• Week 2 (4/4/11- 4/8/11)
• Bias variance analysis, linear classifier - perceptron (slides)
• Linear Discriminant Analysis (old slides: dimension reduction view; new notes )
  

• Week 3 (4/11/11-4/15/11)
• Logistic regression (old slides); notes
• Generative classifier, Bayes and Naive Bayes Classifier (old slides)
  
• Decision Trees (slides) updated 4/18/11 (bug corrected)

• Week 4 (4/18/11 - 4/22/11)
• Neural networks ( slides)
• Nearest Neighbors ( slides)
• Linear SVM (slides). Reading: Bishop PRML Appendix E
  
• Week 5 (4/25/11- 4/29/11)
• SVM Part 2 (slides)
• Computational Learning Theory (slides)
  

• Week 6 (5/2/11 - 5/6/11)
• Ensemble methods (slides updated 5/2/11 )
  
• Midterm (Wed 5/4/11) sample midterm ; Midterm solution
• Model Selection and Regularization(slides) slightly updated on 5/12/11

• Week 7 (5/9/11 - 5/13/11)
• Unsupervised learning(slides: including HAC, Kmeans, Mixture of Gaussian)

• Week 8 (5/16/11 - 5/20/11)
• The Expectation Maximization Algorithm(No slides, see board for notes, which roughly follows Andrew Ng's lecture notes on EM)
• Spectral Clustering( slides)
• Evaluating Clustering and Model Selection( slides)
• Dimension reduction: LDA( slides)

• Week 9 (5/23/11 - 5/27/11)
• Dimension reduction: PCA, ISOMAP( slides)
• Semi-supervised learning: generative model and cotraining( slides)
• Semi-supervised learning: graph based and S3VM (slides)

• Week 10 (5/30/11 - 6/3/11)

• Final exam (6/7/11 6:00PM) Examples: 2009 final (Hand witten solution; a few sample questions

---