TrainingsMasterEconomicsCoursesAutomatic model selection methods

Master EconomicsUE Automatic model selection methods

Content

The objective of this course is to introduce quantitative methods allowing to reduce information. These methods cover different fields of statistics and are based on classical econometric methods (OLS, MLE) or classificatory or principal component methods. The ultimate goal is to study methods to do automatic variable selection in large-scale problems and to apply them to real data.

Course outline :

  • Classification methods
  • Economic factor models
  • Statistical factor models
  • Lasso methods
  • The so-called « General to Specific » method (Hendry, Gets or Autometrics Methodology)

Professional skills

  • Understanding new methods
  • Application on real data
  • Learning new tools or econometric softwares dedicated to the reduction of information

Language used

Main language used by this course: Anglais.

Bibliography

  • Doornik, J.A. and Hendry, D.F. (2015). Statistical model selection with “Big Data”, Cogent Economics & Finance, vol 3, n°1, 1-15.
  • Hendry, D.F. and Doornik, J.A. (2014). Empirical Model Discovery and Theory Evaluation. Automatic Selection Methods in Econometrics. The MIT Press.
  • Richard A. Johnson and Dean W. Wichern, Applied Multivariate Statistical Analysis, Pearson.

Fundamental prerequisites

  • Advanced statistics
  • Introduction to econometrics

Structure and organisation

Lecture. Evaluation based on a group project. Writing of a final document and oral presentation.

Volume of teachings

  • Lectures: 24 hours

The trainings which use this course