An enhanced LGSA-SVM for S&P 500 index forecast

The S&P 500 index is an important representative of worlds' financial market and is influenced by various economic factors. There is a call for automatically select antecedents of S&P 500 index's change in the fast-changing world economy. This paper proposes an enhanced GSA model named LGSA to solve the feature selection and parameter optimization of SVM models for the S&P 500 index movement prediction. The results show that the accuracy of LGSA-SVM model surpasses benchmark SVM, PSO-SVM and GA-SVM model. And the proposed approach could hopefully be adopted for other financial data series automatic forecasting.