摘要
为了找到负荷值与各种影响负荷预测精度因素之间的关系来进行缺损数据处理,提出一个基于关联分类技术的短期负荷数据缺损处理模型。该模型首先对负荷信息系统应用数据规约方法得到规约集,然后利用关联分类算法挖掘出隐含在其中的有趣的满足用户指定的最小支持度和最小信任度的强关联规则,最后通过规则匹配对含有缺损数据的记录进行修补,对有问题的数据判断异常。经仿真分析,应用这种新的数据缺损处理策略可以得到更加精确的预测结果。
In this paper, a model based on associative classification technique is presented in order to find the relationship between the load and the factors correlated with load forecasting accuracy, which can be used in incomplete data processing. Firstly, the rule sets are obtained with the rule methods, and then the inherent and interesting strong association rules are gained through the associative classification algorithm. By matching these rules, the model can remedy the missing values and modify the anomalous data. The simulation result shows that the proposed model can obtain much more accurate forecasting value.
出处
《电力系统及其自动化学报》
CSCD
北大核心
2006年第6期83-86,共4页
Proceedings of the CSU-EPSA
关键词
负荷预测
缺损数据
关联分类
关联规则
load forecasting
incomplete data
associative classification
association rule