Elements of Statistical Learning:
Schedule & Associated Material
Lectures
All lectures will be held in 304, Teknikringen 14. Currently this is a provisional
schedule which may be altered slightly to accommodate participants if necessary.
Part I
The first part of the course will cover the first 8 chapters of the book. These chapters
cover the linear world via linear regression and classification;
the building blocks for high-dimensional learning with the introduction
of splines and regularization methods followed by kernel methods and local regression;
issues with model assessment and inference such as the bias and variance
of an estimator, over-fitting and using cross-validation to choose between models.
Date |
Time |
Book Chapter |
Assigned Reading |
Exercises |
Programming |
|
9th March
|
10:00-12:00 |
- Introduction
- Overview of Supervised Learning
|
|
- Tutorial: 2.3, 2.4, 2.5, 2.6
- Homework: 2.9
|
By the 2nd of April please send me ~1 page that contains a description of
- a problem related to your research which can be tackled via supervised learning,
- the techniques you would like to try out on this problem.
If you don't have access to a labelled dataset talk to me and we'll figure
something out!
|
16th March
|
10:00-12:00 |
Linear methods for regression
|
- Pages 43 -- 52
- (Pages 52 -- 55)*
- Pages 57 -- 73
- (Pages 73 -- 79)*
- Pages 79 -- 81
- Lecture2.pdf
|
- Tutorial: 3.5, 3.7, 3.9, 3.28, 3.3(a), 3.6
- Homework: 3.12, 3.29
|
tba
|
23rd March
|
10:00-12:00 |
Linear methods for classification
|
- Pages 101 -- 135
- Lecture3.pdf
- Slides giving a geometric explanation of the KKT conditions
- Slides attempting to explain how constrained optimization problems can be solved via the dual formulation..
feel free to skim over the latter part of sections when things get a bit technical!
- Tutorial: 4.1, 4.2, 4.6
- Homework: 4.7
|
tba
|
|
30th March
|
10:00-12:00 |
Basis expansion for high-dimensional learning
|
- Pages 139--145
- Pages 151--156
- (Pages 157--161)*
- Pages 161--162
- Lecture4.pdf
The latter part of the lecture I will devote to section 5.8. This section is quite technical
so I suggest that only those who are really keen or have some background in the area stay for
this part of the lecture.
|
- Tutorial: 5.1, 5.7, 5.15, 5.16, 5.17
- Homework: 5.11, 5.13
|
tba
|
13th April
|
10:00-12:00 |
Kernel smoothing methods
|
- Sections 6.1, 6.2, 6.3, 6.5, 6.6, 6.7, 6.8
- Tutorial: 6.1, 6.7, 6.9
- Homework: 6.5
|
tba
|
|
20th April
|
10:00-12:00 |
Model assessment and selection
|
- Section 7.2, 7.3
- Section 7.7, 7.8
- Sections 7.10, 7.11
- Lecture6.pdf
|
- Tutorial: 7.4, 7.6, 7.10
- Homework: 7.5
|
tba
|
27th April
|
14:00-16:00 |
Model inference and averaging
|
- Tutorial: 8.1, 8.4
- Homework: 8.2
|
tba
|
|
* The pages in brackets correspond to (sub)sections which perhaps should only be
read by the keen. (i.e. they're optional)!
This is my record of the
Lectures you've attended. It has gaps! Especially from lecture 5. Please e-mail me if
I have made an error.
Part II
The second part of the course will focus on Structured methods for supervised
learning and the practicalities of working with high dimensional data.
Date |
Time |
Book Chapter |
Assigned Reading |
Exercises |
Programming |
|
6th Nov
|
09:00-11:00 |
Additive models, trees and related methods
|
- Sections 9.1, 9.1.1
- Sections 9.2.1 - 9.2.4
- Sections 9.4, 9.5
- Lecture8.pdf
|
|
tba
|
20th Nov
|
10:00-12:00 |
Boosting and additive trees
|
- Sections 10.1 - 10.4
- Sections 10.6, 10.7
- Sections 10.9, 10.10
- Lecture9.pdf
|
|
tba
|
27th Nov
|
10:00-12:00 |
Random forests and Ensemble learning
|
|
- Homework: 15.1, 15.2, 16.3
|
tba
|
4th Dec
|
10:00-12:00 |
Support vector machines and flexible discriminants
|
|
- Homework: 12.1, 12.2, 12.6, 12.10
|
tba
|
11th Dec
|
11:30-13:00 Teknikringen 14 523 |
Neural Networks
|
|
- Homework: 11.2, 11.3, 11.4
|
tba
|
17th Dec
|
14:00-16:00 |
Deep Learning
|
|
tba
|
tba
|
|