关键词: 石油期货, 预测, 时间序列, 最小二乘支持向量机

Abstract: A novel forecasting model of petroleum futures price based on Least Squared Support Vector Machine（LS-SVM） is proposed.The experiment on the prediction of 2 kinds of daily petroleum futures price recorded in New York Mercantile Exchange（NYMEX） is carried out.RBF neural network prediction method is also applied to petroleum futures price time series.The results indicate that the best precision of fitting and forecasting can be obtained with LS-SVM prediction model，and LS-SVM prediction model outperforms RBF network prediction model.

Key words: petroleum futures, prediction, time series, Least Squared Support Vector Machine（LS-SVM）