5. The Modeler Module

This is a description of the Modeler module whose main goal is to produce a surrogate model from a database stored in a TDataServer.

• 5.1. Introduction
• 5.2. The TLinearRegression Class
• 5.3. Chaos polynomial expansion
  • 5.3.1. Nisp in a nutshell
    • 5.3.1.1. The integration method
• 5.4. The kriging method
  • 5.4.1. Construction of a kriging model
    • 5.4.1.1. Example: construction of a simple Kriging model
    • 5.4.1.2. Deterministic trend and bayesian prior
    • 5.4.1.3. Choice of the initial point of the optimisation
  • 5.4.2. Usage of a Kriging model
    • 5.4.2.1. Prediction of a new data set, one-by-one approach
    • 5.4.2.2. Prediction of a new data set, global approach