UCL - Machine Learning : regression, dimensionality reduction and data visualization

[ELEC2870]

[30h + 30h] 5 credits

This course is taught in the 1st semester.

Teacher(s): Michel Verleysen

Language: English

Level: Second

Aims
Main themes
Content and teaching methods
Other information (prerequisite, evaluation (assessment methods), course materials recommended readings, ...)

Aims

1. To understand and to be able to apply machine learning concepts for analyzing data and signals, in particular in the context of regression and prediction problems;
2. To understand and to be able to apply linear and nonlinear techniques for data visualization;
3. To be able to evaluate the performances of these methods through appropriate techniques;
4. To be able to choose between existing machine learning techniques, according to the nature of the data and signals to be analyzed.

Main themes

See summary.

Content and teaching methods

" Linear regression
" Nonlinear regression with Multi-Layer-Perceptrons
" Clustering and vector quantization
" Nonlinear regression with Radial-Basis Function Networks, Kernel regression
" Probabilistic models for Regression
" Ensemble models
" Feature selection
" Model selection
" Principal Component Analysis
" Nonlinear dimensionality reduction and data visualization
" Independent Component Analysis

Other information (prerequisite, evaluation (assessment methods), course materials recommended readings, ...)

The course necessitates only a basic knowledge in linear algebra. In addition to the course itself there are exercise sessions organized on computers, and students must realize d a project aims at applying machine learning techniques in a specific application context. The exam is oral (if the number of students remains limited enough); the project report is evaluated too.