电力系统中长期负荷预测的改进决策树算法

Improved Decision Tree Method for Mid-Long Term Load Forecasting of Power System

以不完备数据分析(ROUSTIDA)补全算法为数据处理前导,将其与属性组合法相结合,提出应用于电力系统中长期负荷预测的改进决策树算法.算例的计算结果表明,该算法弥补了决策树算法的固有缺陷,具有计算速度快、预测精度高、预测误差变化小等优点.尤其在候选属性繁多,原始数据不完整或不准确时,改进算法更具优越性.

Taking ROUSTIDA algorithm as the pre-process and integrating it with attribute-combination method, this paper puts forward an improved decision tree method applied to mid-long term load forecasting of power system. The results of practical examples prove that this method to some extent makes up the limitations of IDS algorithm and has the advantage of fast calculation, high forecasting accuracy and low error variety. In particular, when selecting-attributes are various and resource data are non-integrated or inexact, this method has much more advantages.