Machine Learning from a statistical perspective

Presenter(s): Marco Puts & Piet Daas (CBS), CBS


Date: 11 May 2021

Webinar aims

Show some caveats when using machine learning algorithms and give an insight what it means to use machine learning for statistical purposes.

Webinar learning outcomes

  • Know how to identify problems with machine learning algorithms.

  • Be aware of coverage and representativity of the population in the training set.

  • Understand what is needed to validate a ML model.

Webinar content

  • Correlation vs. Causation

  • Annotations and the asymptotical behavior towards annotated data.

  • Representativity of training sets

  • Bias in classifications

  • Explainable AI

  • Concept Drift

Difficulty level


Prerequisites for the webinar


Further reading and resources

Presentation material


Recording of the webinar