Novel QSAR combination forecast model for insect repellent coupling support vector regression and K-nearest-neighbor

To improve the precision of quantitative structure-activity relationship (qsar) modeling for aromatic carboxylic acid derivatives insect repellent,a novel nonlinear combination forecast model was proposed integrating support vector regression (svr) and k-nearest neighbor (knn): firstly,search optimal kernel function and nonlinearly select molecular descriptors by the rule of minimum mse value using svr. secondly,illuminate the effects of all descriptors on biological activity by multi-round enforcement resistance- selection. thirdly,construct the submodels with predicted values of different knn. then,get the optimal kernel and corresponding retained sub-models through subtle selection. finally,make prediction with leave-one-out (loo) method in the basis of reserved sub-models. compared with previous widely used models,our work shows significant improvement in modeling performance,which demonstrates the superiority of the present combination forecast model.