The COmmon REactivity PAttern (COREPA) approach is a probabilistic classification method formalized specifically to advance the mechanistic QSAR development by addressing the impact of molecular flexibility on stereo electronic properties of chemicals. Similarity between chemicals is analyzed by comparing their conformational distributions. The COREPA-M system automatically identifies the reactivity parameters that best discriminates chemicals in groups defined by comparable activity endpoints. A detailed Bayesian decision tree is then used for classifying untested chemicals with measures of "goodness of fit" criteria. The automated building of decision tree for classification of chemicals provided opportunity for a detailed statistical evaluation of derived model. The probabilistic nature of the COREPA classification supports the concept of a continuous behavior of chemicals because some conformers of the same chemical can be classified in different activity classes. In the initial formulation of COREPA, the conformer distributions of chemicals for only one stereo electronic parameter at a time were analyzed for the series of biologically similar chemicals. The probability distributions of a suite of parameters simultaneously are applied in the new formulation of the method. The multi-dimensional formulation of the COREPA enhanced the reliability of the method to discriminate complex reactivity patterns.