Data mining and statistical learning, 15 credits

Course Leader: Anders Grimvall
Phone: tel. +46-(0)13-28 14 82

Target group

Ph.D. candidates in statistics and other disciplines where knowledge of statistics is essential


Undergraduate courses in probability theory and statistical inference


The course aims to make provide the student with a solid background and understanding of statistical learning and modern regression techniques


Review of linear methods for prediction and classification
Non-linear methods for a single predictor
Model assessment and selection
Model inference
Specific methods for supervised learning and data mining


Hastie, T., Tibshirani, R., Friedman, J. (2001). The Elements of Statistical Learning. New York :Springer


Active participation in classes
Oral presentations of selected chapters from the textbook
Computer exercises


Seminars approximately once a week.

Seminar 1: Linear methods for regression (Chapter 3)

Seminar 2: Linear methods for classification (Chapter 4)

Seminar 3: Smoothing splines and wavelet smoothing (Chapter 5)

Seminar 4: Kernel methods and local regression (Chapter 6)

Seminar 5: Information criteria, effective number of parameters, cross-validation (Chapter 7)

Seminar 6: Bootstrap, the EM algorithm, MCMC (Chapter 8)

Seminar 7: Generalized additive models and regression trees (Chapter 9)

Seminar 8: Boosting and multiple additive regression trees (Chapter 10)

Seminar 9: Projection persuit regression and neural networks (Chapter 11)

All seminars will take place in "Kompakta rummet", Department of mathematics, Building B, entrance 23.