Helsinki University of Technology →
Department of Computer Science and
Laboratory of Computer and Information Science → Teaching →
T-61.3050 Machine Learning: Basic Principles → 2007 → Lectures
The lectures take place in T1 from 11 September to 11 December 2007 on Tuesdays at 10-12. There is no lecture on 30 October.
|Date||Topic||Source Material||Slides||Other Material|
|1||11 September||Introduction||Alpaydin (2004) Ch 1|
|2||18 September||Supervised Learning||Alpaydin (2004) Ch 2|
|3||25 September||Bayesian Decision Theory||Alpaydin (2004) Ch 3|
|4||2 October||Bayesian Networks||Alpaydin (2004) Ch 3&4|
|5||9 October||Model Selection||Alpaydin (2004) Ch 4|
|6||16 October||Multivariate Methods||Alpaydin (2004) Ch 4&5||Alpaydin's slides Ch 5|
|7||23 October||Dimensionality Reduction||Alpaydin (2004) Ch 6||Alpaydin's slides Ch 6|
|30 October||No lecture.|
|8||6 November||Dimensionality Reduction & Clustering||Alpaydin (2004) Ch 6&7|
|9||13 November||Clustering & Decision Trees||Alpaydin (2004) Ch 7&9|
|10||20 November||Decision Trees & Linear Discrimination||Alpaydin (2004) Ch 9&10||Mitchell (2005) Generative and Discriminative Classifiers: Naive Bayes and Logistic Regression|
|11||27 November||Machine learning guest lectures|
The future topics, shown in italics, are preliminary and may still change.
The slides are meant to be a presentational aid for the lectures. They are provided here to make it easier to follow the lectures and make notes. The slides are not a replacement for the text book. We strongly recommend that you also read the book chapters. If you are are not happy with the course slides, you should probably also check out the Alpaydin's own slides.
Please give feedback of the lectures or slides. You can also consider using the newsgroup.
I have given Edita a permission to print the slides to the students on request.
The course uses Alpaydin (2004) as a text book. The other excellent text books mentioned above are however all recommended reading for the interested. In this course they can be used as supplementary material or alternative sources for the course topics.
Bishop (2006) presents the principles and a wide selection of methods and approaches of machine learning from a distinctively Bayesian viewpoint. Many of the chapters are quite nice, but some parts of the book however suffer from hiding the essential principles behind details and long derivations. Duda et al. (2000) is a classic which has a comprehensive review of many methods and principles of machine learning. Mitchell (1997), while already a bit aged, is a good introductory source especially for more learning theoretic and algorithms based approaches, such as concept learning and decision trees. Sivia & Skilling (2000) is one of the most readable and clear introductions to Bayesian data analysis. If you really want understand the Bayesian data analysis, this book would be a good choice; the first edition was what the lecturer used when he was young a long time ago. Vapnik (2000) presents the statistical learning theory for those who like it.
Some source code used to generate various images in the slides (somewhat undocumented, provided as is):
Page maintained by email@example.com, last updated Wednesday, 02-Jan-2008 12:38:12 EET