Fuzzy Filtering for Environmental Behavior Modeling of Chemicals

Toxicity Modeling

A fundamental concern in the Quantitative Structure-Activity Relationship approach to toxicity evaluation is the generalization of the model over a wide range of compounds. The data driven modeling of toxicity, due to the complex and ill-defined nature of eco-toxicological systems, is an uncertain process. The development of a toxicity predicting model without considering uncertainties may produce a model with a low generalization performance. This work presented a novel approach to toxicity modeling that handles the involved uncertainties using a fuzzy filter, and thus improves the generalization capability of the model. The method was illustrated by considering a data set built up by U.S. Environmental Protection Agency referring to acute toxicity 96-h $LC_{50}$ in the fathead minnow fish (Pimephales promelas). The data set contains 568 compounds representing several chemical classes and modes of action.

Bioconcentration Factor Modeling

This work presented a fuzzy filtering based technique for rendering robustness to the modeling methods. A case study, dealing with the development of a model for predicting the bioconcentration factor (BCF) of chemicals, was considered. The conventional neural/fuzzy BCF models, due to the involved uncertainties, may have a poor generalization performance (i.e. poor prediction performance for new chemicals). The approach to improve the generalization performance of neural/fuzzy BCF models consists of 1) exploiting a fuzzy filter to filter out the uncertainties from the modeling problem, 2) utilizing the information about uncertainties, being provided by the fuzzy filter, for the identification of robust BCF models with an increased generalization performance. The approach was illustrated with a data set of 511 chemicals taking different types of neural/fuzzy modeling techniques.

Related Publications