Statistical Machine Learning - Problem solving sessions

All material for the problem solving sessions will be provided on this page

Each pen-and-paper class:

One pdf file with all problems and solutions

Each computer class:

Jupyter notebook with problems.
Direct link to run notebook in Google Colab.
Notebook exported to HTML with solutions.

Data used in the computer classes can be downloaded directly in the notebooks. You can also download all datasets in a single zip-file.

Scheduling

Each problem solving session is given two times with two sessions in parallel. In the schedule, we have marked which session is given with a comment in the form of L followed by the session number as listed below. For example, the first problem solving session has two time slots commented with "L1" and each time slot has two rooms booked. You are free to visit the session that suits you the best.

For the computer based sessions you are mainly supposed to work on your own laptop. If you do not want to bring your own computer, be sure to go to the session booked in a computer room.

Material

If you find any typos or errors in the problems or solutions, please let us know! We update the material continuously during the course.

	Topic	Type	Recommended problems	Additional
1	Linear regression	Pen-and-paper	1, 2, 3	4, 5
2	Linear regression	Computer class	1, 2, 3, 4	5
3	Logistic regression, LDA, QDA, kNN	Pen-and-paper	1, 2, 3, 5	4, 6, 7
4	Logistic regression, LDA, QDA, kNN	Computer class	1, 2, 3, 4	5
5	Bias and variance, model selection, cross validation	Pen-and-paper	1, 2, 3, 4	5, 6
6	Bias and variance, model selection, cross validation	Computer class	1, 2, 3	4
7	Tree-based methods	Pen-and-paper	1, 2, 3, 4
8	Tree-based methods	Computer class	1, 2	3
9	Boosting	Computer class	1, 2, 3	4, 5
10	Neural networks	Pen-and-paper	1, 2, 3, 4