Fragmental descriptors in (Q)SAR: prediction of the assignment of organic compounds to pharmacological groups using the support vector machine approach

Abstract

The structure—activity classification models for prediction of the assignment of organic compounds to 40 pharmacological groups were constructed in the framework of the fragmental approach using the support vector machine technique, the Platt—Wu probabilistic model, and resampling procedure. The models constructed allow one to predict possible types of the side pharmacological effects of drugs.