Méthodes de classification
UE-EIG.00121

Teacher(s): Donzé Laurent
Level: Master
Type of lesson: Lecture
ECTS: 4.5
Language(s): French
Semester(s): SS-2025

The course is an introduction to classification methods:

  1. Discriminant analysis (Introduction; binary classification; multiclass classification)
  2. Splines (Bézier curves and splines; splines smoothers)

  3. Generalized additive models and regression trees (Generalized additive models; classification and regression trees; MARS; boosting algorithms)
  4. Random forests and transformation forests

The theoretical concepts and the methods are illustrated by practical applications. R is the software used.

Moodle


Training aims

The course is part of the supplied of Master courses in applied statistics, which is a packet of four half-yearly ones of 4.5 ECTS (one by semester) over two years:

  1. Topics in multivariate statistics
  2. Introduction to Bayesian statistics
  3. Inference, evaluation and selection of models
  4. Classification methods

Although they complete each other, they can be chosen separately. Of general interest, the set of courses gives a broad view of problems and applied statistical methods, and of the data science. A server of Jupyter notebooks completes the course.

There is no specific public. Although the courses are primarily conceived for students of the Faculty of Management, Economics and Social Sciences, they can be attended by other students. 

The student will benefit not only of theoretical knowledge but is also skilled in the use of the methods presented during the course.


Documentation

A script with a listing of references is provided. The student will find on the course's Moodle platform other resources.