论文部分内容阅读
在推导加权最小二乘支持向量机数学模型的基础上,基于启发式学习算法并结合滚动窗的思想,提出基于滚动窗法最小二乘支持向量机的稳健预测模型.为了缩短模型的预测运行时间,将启发式算法进行改进后,采用迭代求逆方法,在不丧失预测精度的基础上,很大程度地缩短预测时间.最后通过仿真实例验证这个模型可以成功抑制奇异点,实现稳健预测并取得良好效果.
Based on the mathematical model of weighted least square support vector machines and based on the heuristic learning algorithm combined with rolling windows, a robust prediction model based on rolling window least squares support vector machine is proposed.In order to shorten the prediction run time , The heuristic algorithm is improved, and the iterative inversion method is used to shorten the prediction time greatly without losing the prediction accuracy.At last, the simulation example shows that this model can successfully suppress the singularities and achieve robust prediction Good effect.