Statistical Data Mining and Machine Learning

Statistical Data Mining and Machine Learning (SC4 HT2015)

Term:	Hilary Term, January 18 - March 14, 2015
Lecturer:	Dino Sejdinovic
Teaching Assistants:	Part C: Owen Thomas and Helmut Pitters MSc: Konstantina Palla and Maria Lomeli
Lectures:	1400-1500 Mondays in Math Institute L4 (weeks 1-4,6-8), L3 (week 5) 1000-1100 Wednesdays in Math Institute L3
Problem Sheets: Part C Classes: MSc Classes:	Due 1000 the Mondays prior to classes in 1 South Parks Road (SPR). 1500-1700 Tuesdays (Weeks 3-8) in 1 SPR Seminar Room. 1500-1700 Wednesdays (Weeks 3,5,7,9) in 1 SPR Seminar Room.
Practicals: (MSc only)	1400-1800 Friday Week 4 (unassessed) in 1 SPR Computing Lab. 1400-1800 Friday Week 8 (assessed) in 1 SPR Computing Lab.

Course Materials

Lecture 1: Intro, Exploratory Data Analysis, PCA
[slides], [4 slides per page]
Lecture 2: SVD and Biplots, MDS
[slides], [4 slides per page]
Lecture 3: K-means, Hierarchical Clustering
[slides], [4 slides per page]
Lecture 4: Mixture Models, EM Algorithm
[slides], [4 slides per page]
Lecture 5: Loss and Risk, LDA
[slides], [4 slides per page]
Lecture 6: FLDA, QDA, Naive Bayes
[slides], [4 slides per page]
Lecture 7: Generative vs Discriminative Learning, Logistic Regression, Performance Measures
[slides], [4 slides per page]
Lecture 8: Convex Optimization, Support Vector Machines
[slides], [4 slides per page]
Lecture 9: Overfitting and Regularization
[slides], [4 slides per page], [rnw source]
Lecture 10: Kernel Methods
[slides], [4 slides per page]
Lecture 11: Nearest Neighbours, Neural Networks
[slides], [4 slides per page]
Lecture 12: Model Complexity and Validation
[slides], [4 slides per page]
Lecture 13: Decision Trees and Bagging
[slides], [4 slides per page]
Lecture 14: Random Forests and other Ensemble Methods, Boosting
[slides], [4 slides per page]
Lecture 15: Bayesian Learning
[slides], [4 slides per page]
Lecture 16: Bayesian Nonparametrics and Gaussian Processes
[slides], [4 slides per page]

You will need to use the R statistical programming language and environment for this course.

An introduction to R.
[pdf]
If you do not have access to a computer with R already installed, you can download it at
http://cran.r-project.org/.

Recommended textbooks on statistical data mining and machine learning:

Hastie, Tibshirani and Friedman, The Elements of Statistical Learning, Springer-Verlag.
[ebook]
Ripley, Pattern Recognition and Neural Networks, Cambridge University Press.
James, Witten, Hastie and Tibshirani, An Introduction to Statistical Learning with Applications in R, Springer-Verlag.
[ebook]
Bishop, Pattern Recognition and Machine Learning, Springer.
Duda, Hart and Stork, Pattern Classification, Wiley-Interscience.

Background Review Aids: