基于降维聚类技术的电力负荷数据挖掘研究

随着电力网络的高速发展，电力负荷数据的规模与维数急速增长。为了分析数据背后的有效信息，可以采用聚类分析的手段对电力负荷数据进行挖掘分析，为异常用户检测、能效管理提供有效的应用价值。根据美国能源信息网获得的实验数据集，利用降维算法对预处理后的数据进行降维分析，分析出不同维度下5种降维算法的降维效果，然后选择KPCA和ISOMAP降维技术与K-means聚类分析算法进行结合，比较组合算法与单独K-means算法的聚类精度与聚类效率，得出结合降维技术，可以有效提高聚类分析算法的聚类能力。

Research on power load data mining based on dimensionality reduction clustering technology

With the rapid development of power network，the scale and dimension of power load data are increasing rapidly.In order to analyze the effective information behind the data，cluster analysis can be used to mine and analyze the power load data，so as to provide effective application value for abnormal user detection and energy efficiency management.According to the experimental data set obtained by the American energy information network，the dimensionality reduction algorithm is used to reduce the dimensionality of the preprocessed data，and the dimensionality reduction effects of five dimensionality reduction algorithms under different dimensions are analyzed.Then KPCA and ISOMAP dimensionality reduction technology are combined with K-means clustering analysis algorithm to compare the clustering accuracy and clustering efficiency of combined algorithm and single k-means algorithm.It is concluded that the combination of dimension reduction technology can effectively improve the clustering ability of clustering analysis algorithm.