数据挖掘与非正常日的负荷预测

提高非正常日的负荷预测精度是当前负荷预测工作的难点。文中提出了一种基于知识库的事先判别突变并做出适当处理的预测流程，介绍了利用数据挖掘的决策树技术建立知识库的方法，并给出了几种典型的非正常日修正模型。最后，通过对长时期负荷预测数据的统计分析，说明了新方法的有效性和实用性。

DATA MINING AND SHORT-TERM LOAD FORECASTING FOR ABNORMAL DAYS

The current difficulty of short-term load forecasting is the forecasting of abnormal days' load. This paper puts forward a forecasting flow based on the knowledge bases, to prefigure the abnormity and make proper treatment. It also introduces the method of constructing the knowledge base using decision tree technique of data mining, and gives several practical amending models of typical abnormal days. In the end, it illuminates the new'method's effectiveness and practicability through the statistical analysis of the load forecasting data over a long period of time.