基于混合高斯模型的用电量计量数据聚类算法研究

Research on clustering algorithm of electricity consumption measurement data based on hybrid Gaussian model

针对传统用电量计量数据聚类算法中存在聚类性能较差的问题,提出一种基于混合高斯模型的用电量计量数据聚类算法。通过主成分分析方法对用电量计量数据进行线性降维处理,利用线性降维后的用电量计量数据实施数据预处理,具体步骤包括用电量负荷数据矩阵表示、异常用电量数据修正及辨识以及用电量数据归一化处理,根据预处理后的用电量计量数据,采用混合高斯模型实现用电量计量数据聚类。为了证明基于混合高斯模型的用电量计量数据聚类算法的集中聚类性能较强,将传统用电量计量数据聚类算法与该算法进行对比实验,实验结果证明该算法的集中聚类性能优于传统用电量计量数据聚类算法,更适用于用电量计量数据的聚类。

Aiming at the problem that the clustering performance is poor in the traditional power consumption measurement data clustering algorithm,a power consumption measurement data clustering algorithm based on a mixed Gaussian model is proposed.the method comprises the following steps of:carrying out linear dimension reduction processing on the power consumption measurement data by a main component analysis method,carrying out data preprocessing by utilizing the power consumption measurement data after linear degradation,The abnormal power consumption data correction and identification and the power consumption data normalization processing are carried out,and the power consumption measurement data clustering is realized by adopting a mixed Gaussian model according to the pre⁃processed power consumption measurement data.In order to prove the power consumption measurement based on the mixed Gaussian model according to the clustering algorithm,the clustering algorithm of the traditional power consumption is compared with the algorithm,and the experimental results show that the concentrated clustering performance of the algorithm is better than that of the traditional power consumption measurement data clustering algorithm,and is more suitable for clustering the power consumption measurement data.