Application of Support Vector Machine to Capital Flow Risks Prediction

Abstract

Under the opening economic circumstances, forecasting the risks of capital flow has special significance. For effectively early warning the risks associated with capital flow, this study applies support vector machine (SVM) to the domain of capital flow in an attempt to suggest a new model with better explanatory power and stability. To serve this purpose, a grid-search technique using 5-fold cross-validation is used to find out the best parameter value of kernel function of SVM. In addition, to evaluate the prediction accuracy of SVM, this study compares its performance with that of three-layer fully connected back-propagation neural networks (BPNs). The experiment results show that SVM outperforms the BPNs.