Statistical Principles in Machine Learning for Small Biomedical Data
Date: Tuesday 10 December 2024, 13:00-16:00
Room: Prolog (room 2465), Ole-Johan Dahls hus (OJD)
Instructors: Manuela Zucknick (main), Theophilus Asenso
Welcome!
- The goal of the workshop is to introduce key connected concepts in machine learning, such as generalisation, overfitting, bias-variance trade-off and regularisation.
- The workshop is intended for students and researchers who are interested in applying machine learning methods to small data (few samples, but potentially many features) or noisy data (e.g. biomedical data)
- Workshop material can be found in the workshop github repository.
Learning Objectives
At the end of the tutorial, participants will be able to
- understand key concepts for training machine learning models on small or noisy data, such as generalisation, overfitting, bias-variance trade-off and regularisation;
- understand how to incorporate data structure in the regularisation process.
Pre-requisites
- Basic familiarity with R
- Introductory level statistics, including regression
Schedule
| Time | Topic | Presenter |
|---|---|---|
| Now | Preparations | |
| 13:00 - 14:00 | (Supervised) machine learning with small data | Manuela Zucknick |
| R lab 1 | Manuela Zucknick | |
| 14:15 - 15:15 | Overfitting, regularisation and all that | Manuela Zucknick |
| R lab 2 | Manuela Zucknick | |
| 15:30 - 16:00 | Hierarchical models and structured penalties | Theophilus Asenso |