Prediction of melting points of a diverse chemical set using fuzzy regression tree

Abstract

The classification and regression trees (CART) possess the advantage of being able to handle<br />large data sets and yield readily interpretable models. In spite to these advantages, they are also<br />recognized as highly unstable classifiers with respect to minor perturbations in the training data.<br />In the other words methods present high variance. Fuzzy logic brings in an improvement in these<br />aspects due to the elasticity of fuzzy sets formalism. ACS, which is a meta-heuristic algorithm<br />and derived from the observation of real ants, was used to optimize fuzzy parameters. The<br />purpose of this study was to explore the use of fuzzy regression tree (RT) for modeling of<br />melting points of a large variety of chemical compounds. To test the ability of the resulted tree, a<br />set of approximately 4173 structures and their melting points were used (3000 compounds as<br />training set and 1173 as validation set). Further, an external test set contains of 277 drugs were<br />used to validate the prediction ability of the tree. Comparison the results obtained from both trees<br />showed that the fuzzy RT performs better than that produced by recursive partitioning procedure.