First semester
Linear Models
- Teacher(s)
- Basile DE LOYNES
- Course type
- STATISTICS
- Correspondant
- Basile DE LOYNES
- Unit
-
Module 2-01: Statistical modeling
- Number of ECTS
- 3
- Course code
- 2ASTA01
- Distribution of courses
-
Heures de cours : 15
Heures de TP : 18
- Language of teaching
- French
Objectives
Understand the construction of a linear regression model
Carry out a complete linear regression study
Perform a constructive residual analysis
Select relevant variables
Use R software packages and the statsmodel library under Python
Course outline
Reminders: orthogonal projections, Gaussian vectors and conditional expectation
The multiple linear regression model
The multiple linear regression model under Gaussian assumption
Model validation, variable selection, diagnostic measures
Prerequisites
Standard probability laws, Gaussian vectors, confidence regions, statistical tests, matrix algebra, orthogonal projection