A Constraint Projection and Genetic Algorithm Based Support Vector Machines Selective Ensemble

This paper proposes a novel selective ensemble algorithm of support vector machines based on constraint projection and genetic optimization. Firstly, projective matrices are determined upon randomly selected must-link and cannot-link constraint sets, with which original training samples are transformed into different representation spaces to train a group of base classifiers. Then, genetic algorithm is utilized to learn the optimal weighting factors to combine them effectively. Experimental results on UCI datasets show that the proposed algorithm improves generalization performance of support vector machines significantly, which outperforms classical ensemble algorithms, such as Bagging, Boosting, feature Bagging and LoBag.