时间相关性经验知识与SVM的融合方法研究

Incorporating Method of Time Correlation and Support Vector Machine

时序数据在时间维度上存在着很强的时间相关性,在时序预测中,利用时序数据的时间相关性特点,构造了一种适用于时序数据预测的时序核函数,实现了将时间相关性融合于支持向量机,并通过人工数据和真实数据验证了时序核函数解决时序预测问题的有效性,并与传统核函数相比具有较好的泛化能力。

It is well known that there is high time correlation among time series data.This paper presented a new kernel function adapted in prediction for time series data,which incorporated the time correlation of time series into Support Vector Machine（SVM）.Simulation experiments were performed to test the robustness of the findings on artificial stimuli and field data,which demonstrated that the presented kernel function can help to improve the fitting effect and obtain better generalization performance by comparing with the traditional kernel function of SVM.