基于自组织映射支持向量机的日前电价预测

Day-ahead Electricity Price Forecasting Using Support Vector Machines Based on Self-Organizing Map

针对电力市场中日前24点电价特性差异较大、采用单一模型很难描述的特点,建立多个模型分别对其进行预测,将数据空间按时点划分成24个子空间,然后根据这些子空间的相似性通过自组织映射对其进行自动聚类,并在不同类别的子空间分别建立支持向量机模型进行训练和预测。应用上述方法对PJM电力市场2005年8月的31天日前24点电价进行预测,结果表明该方法能够有效提高预测精度。

According to the fact that in electricity market the features of electricity prices at day-ahead 24 points in time differ greatly and it is hard to describe these features by a single model, so multi-models are built to forecast them respectively. In accordance with points in time the data space is divided into 24 sub-spaces, then in the light of the similarity of these sub-spaces, they are automatically clustered by means of self-organizing mapping （SOM） and the support vector machine （SVM） models are established in different classified sub-spaces to perform training and forecasting. Using this method, the electricity prices at day-ahead 24 points of PJM electricity market on Aug. 3 1, 2005 are forecasted. Forecasting results show that the proposed method can improve forecasting accuracy effectively.