基于粗糙集改进的支持向量机模型及其在短期电力负荷预测中的应用

Study of Short-term Load Forecasting Based on Improved Support Vector Machine with Rough Set

通过对传统的最小二乘支持向量机模型和粗糙集理论的研究,提出了一种基于粗糙集理论进行改进的最小二乘支持向量机预测技术,将粗糙集原理的属性约筒与特征提取技术运用到输入指标的选取上,保留有用信息并剔除无用信息。最后,以美国PJM市场2012年1月至9月的日24点历史负荷为算例,对该时间段电力负荷进行模拟仿真。结果表明,经过粗糙集属性约简改进后的LS-SVM预测模型大大提高了其预测精度,拟合效果显著提高。

This paper studies the traditional least squares support vector machine model and the rough set theory and puts forward an improved model that is Improved Support Vector Machine based on rough set. The new model is characterized by rough set theory attribute reduction and feature extraction technology is applied in the prediction of the reasonable structure of sample set and get the useful information reserved, the useless information almost eliminated and reach the best fitting and forecast purposes. At last, this paper applied the improved model to short-term load. The results show the improved model＇s accuracy is higher.