KPCA SVM with GA Model for Technological Achievements of College Students Forecasting

Abstract

A two-stage neural network architecture constructed by combining Support Vector Machines (SVM) with kernel principal component analysis (KPCA) and genetic algorithms (GAs) is proposed for technological achievements of college students forecasting. In the first stage, KPCA is used as feature extraction. In the second stage, KPCA SVM is used to regression estimation by finding the most appropriate kernel function and the optimal learning parameters with GAs. By examining the technological achievements data, it is shown that the proposed method achieves both significantly higher prediction performance and faster convergence speed in comparison with a single SVM model. And KPCA SVM outperforms principal component analysis (PCA) SVM.