Exchange Rates Forecasting with Least Squares Support Vector Machine

Abstract

A novel forecasting model of foreign exchange market based on least squares support vector machine (LS-SVM) is proposed in this paper. The experiment on the prediction of four kinds of daily exchange rate recorded is carried out. Grid search method is used to determine the LS-SVM parameters automatically in the forecasting process. The results show the precision of fitting and forecasting are very high, which indicates that LS-SVM is a feasible and valid approach for forecasting exchange rate time series.