基于混沌理论和支持向量机的电价预测

根据电价时间序列的混沌特性,结合混沌理论和支持向量机方法提出了一种新的电价预测模型。该模型基于混沌理论对电价时间序列进行相空间重构,并根据相空间演变规律确定模型的输入输出结构,然后采用支持向量机拟合相点演化的非线性关系。为增强模型的泛化推理能力,训练样本按照预测相点最近邻点原理选择。对美国PJM电力市场边际电价历史数据的仿真研究表明,文中提出的预测模型能有效、稳定地提高电价预测精度。

Electricity price forecast based on chaotic theory and support vector machine

According to the chaotic characteristics of the electricity price time series,a new model of electricity price forecast based on the chaotic theory and the support vector machine(SVM) is presented.In this model,the phase space of time series electricity price data was restructured based on the chaotic theory,the input and output of the model were determined in terms of the evolvement law of the phase space,and the SVM was used for fitting the nonlinear development of phase points.To enhance the generalization ability of the forecasting model,the nearest neighboring points of the forecasting phase points were chosen as training samples.The historical data from the PJM market was used in the case study to forecast the marginal price.The simulation shows that the proposed model can effectively and stably improve the electricity price forecast precision.