You are here: Home News Seminar: Model Selection with Applications in the Life Sciences

Seminar: Model Selection with Applications in the Life Sciences


In mechanistic modeling and model-driven data analysis of biological systems, the choice of the mathematical model is fundamental: all statistical analysis relies on the assumption that the true mathematical model is known and the data is generated from that model.

In practice, the biological mechanism generating the data is often not fully known, and instead a set of biological hypotheses, translating into a set of candidate models, are at hand. The topic of model selection deals with the question of choosing an appropriate model for given data.

Due to the importance of this question, there are numerous approaches for model selection stemming from all different branches of statistics and data analysis, e.g. approaches based on information theory, like the Akaike Information Criterion, frequentist approaches like likelihood-ratio tests, Bayesian approaches like Bayes factors, or approaches from machine learning like cross-validation or regularization.

In this seminar, we aim to get an overview and understand the rationale of the major approaches, and their differences, and see their applications in the life sciences.

When: Thursday, 16:15-17:45


Document Actions