Computational Intelligence and Predictive Toxicology

In this contribution we present investigations about the use of computational intelligence for toxicity prediction of pesticides. Therefore different molecular descriptors are computed and the correlation behavior of the different descriptors in the descriptor space is studied. In a first step 164 pesticides are considered and 175 descriptors are taken into account. From this set of data preliminary results using feed-forward artificial neural networks, namely multi-layer perceptrons, and B-spline networks as neuro-fuzzy system are given.