Classification of Skin Sensitizers on the Basis of Their Effective Concentration 3 Values by Using Adaptive Boosting Method

A new quantitative structure-activity relationship (QSAR) model for the classification of 161 skin sensitizers has been developed with adaptive boosting (Adaboost). The selection of variables for each descriptor was performed with particle swarm optimization (PSO). Among all descriptors in the model, the Radial Distribution Function+3DMolecular Representation of Structure based on Electron diffraction(RDF+3D-MoRSE) descriptor exhibited the highest accuracy in the predictions. On a known compound data set, the AdaBoost model improved the prediction accuracy of the training set and the test set to 100.0% and 89.8%, up from 100.0% and 85.7% when only the support vector machine (SVM) was applied.