The Hybrid Dynamic Prototype Construction and Parameter Optimization with Genetic Algorithm for Support Vector Machine

The optimized hybrid artificial intelligence model is a potential tool to deal with construction engineering and management problems. Support vector machine (SVM) has achieved excellent performance in a wide variety of applications. Nevertheless, how to effectively reduce the training complexity for SVM is still a serious challenge. In this paper, a novel order-independent approach for instance selection, called the dynamic condensed nearest neighbor (DCNN) rule, is proposed to adaptively construct prototypes in the training dataset and to reduce the redundant or noisy instances in a classification process for the SVM. Furthermore, a hybrid model based on the genetic algorithm (GA) is proposed to simultaneously optimize the prototype construction and the SVM kernel parameters setting to enhance the classification accuracy. Several UCI benchmark datasets are considered to compare the proposed hybrid GA-DCNN-SVM approach with the previously published GA-based method. The experimental results illustrate that the proposed hybrid model outperforms the existing method and effectively improves the classification performance for the SVM.