Modeling

Modeling

    • Select modelling technique
    • Generate test design
    • Build model
    • Assess model


Prepared data are suitable inputs for modelling algorithms. Different techniques are used: decision trees, regressions, neural networks, link analyses, Bayes-learning, clustering, etc. This is also the time to evaluate the quality and validity of our model, in case a directed learning algorithm was chosen. Based on this, we create training and a test database. The most essential validation procedures are the following:

    • Simple validation
    • Cross-validation
    • Bootstrapping