本文在结构风险最小化的准则下,从提高样本外套期保值效率的视角,建立了基于支持向量机的套期保值新模型,并利用我国沪深300股票指数和沪深300股指期货仿真交易的历史数据进行了实证检验,并与基于最小二乘回归的套期保值模型进行了对比分析。实证结果表明本文提出的新套期保值技术能够有效提高样本外的保值效果,且该方法具有良好的鲁棒性,从而具有较好的理论和应用价值。
From the view of improving the hedge efficiency of out-of-sample, this paper propose a new hedge model based on the SVM under the criterion of structure risk minimization, and use the historical data of Hu - Shen 300 index and Hu - Shen stock index future simulation trade to empirical test, and compare the model with the hedge model based the OLS regression method. The result suggests that our new model with good robustness can improve the hedge efficiency of out - of - sample effectively, so it has fine theoretical and utilization value.