First order models for the Predictive ToxicologyChallenge 2001

This paper discusses the \Leuven" submission 1 to the Predictive Toxicology Challenge 2001. A brief account of some preparatory work is given, followed by a more detailed description of the approach that in the end led to the submitted model, and of the model itself. Both the approach and the model are evaluated from a data mining point of view, and a number of conclusions are drawn.