Why a linear model?

A linear model, in the broadest sense, is the basis for many more complex models, which are used to make predictions of a quantity of interest. This course provides the necessary elements to master this essential model, with a view to its use in machine learning and to create metamodels.


2-day course


  • Learn about the different types of metamodels.
  • Know how to build and exploit a metamodel by linear regression.
  • Know the main designs of experiments and methods of validation


  • Good knowledge of probability theory.
  • Knowledge of Python recommended.


  • R&D Engineer
  • Design Office Engineer
  • Operational safety Engineer.

Upcoming sessions

  • May 20-21, 2021
  • November 4-5, 2021


Paris (12ème ou 7ème arr. suivant disponibilité)




Introduction to metamodels

Design of experiments

Choosing the model

Validating a metamodel

Linear regression model

  • Asymptotic formulation and law of estimators
  • Validation and statistical tests

Practice work


Application form

Register or ask for further information. Personalised training courses can also be provided. The location of the training is subject to change. Training courses can be held at a distance. In this case, the training can be delivered in several short sessions (e.g. 2 hours) via videoconference software, spread over several days, in agreement with the trainer and the participants.


The following information is retrieved for the following purpose: management of the commercial relationship.
Phimeca undertakes to comply with the regulations in force applicable to the processing of personal data and, in particular, Regulation (EU) 2016/679 of the European Parliament of 27 April 2016 applicable from 25 May 2018 (the “European Data Protection Regulation” or GDPR).
You have the right of withdrawal and can write to dpo@phimeca.com