基于GRU-MMD的电力数据挖掘方法

Power Data Mining Method Based on GRU-MMD

针对如何提高含有不同特征的电力数据挖掘效率问题,提出了基于门控循环单元(GRU)-最大均值差异(MMD)的电力数据挖掘方法。采用GRU方法对电力数据进行挖掘,得到短期电力负荷的预测值。当电力数据中源领域数据与目标领域数据特征相差较大的时候,为了提高传统的GRU挖掘效率,提出了采用MMD方法计算源领域和目标领域之间的差异,根据差异值调整GRU网络结构,从而提高挖掘效果。仿真结果分析,验证了经过MMD处理的BP,LSTM,GRU方法负荷预测精度均高于传统BP,LSTM,GRU算法,且本文所提的GRU-MMD算法具有更高的预测精度,验证了本文所提方法的可靠性。

In order to improve the efficiency of power data mining with different characteristics,a gated recurrent unit(GRU)-maximum mean difference(MMD)based power data mining method is proposed.The GRU method is adopted to mine the power data,and the predicted value of short-term power load is obtained.In order to improve the efficiency of traditional GRU mining,MMD method is adopted to calculate the difference between the source field and the target field when the characteristics of the data in the power data are quite different from those in the target field,and the GRU network structure is adjusted according to the difference,so as to improve the mining effect.Simulation result analysis verifies that the load prediction accuracy of BP,LSTM and GRU methods treated by MMD is higher than that of traditional BP,LSTM and GRU algorithms.Moreover,the GRU-MMD algorithm proposed in this paper has higher prediction accuracy,which verifies the reliability of the proposed method.