Power Consumption Portrait of Users Based on Improved ISODATA Clustering Algorithm

With the deepening of information construction and the rapid development of power business, residents’ electricity consumption behaviors show different characteristics. Power grid enterprises have accumulated rich and valuable data resources, and the analysis of users’ electricity consumption experience and corresponding users’ behaviors has been paid more and more attention. That is, it is necessary to draw a picture of the electricity consumption from the electricity consumption characteristics in combination with the electricity consumption information of users. The realization of process is mostly realized by clustering. Aiming at the problem that traditional K-means algorithm is sensitive to the initial clustering center, this paper proposes an improved clustering analysis method of power consumption load based on ISODATA. When the cluster center is unknown, the clustering effect of this method is less affected by the initial center, and the clustering measurement parameters can be dynamically calculated to achieve more effective data clustering. Finally, based on the actual data, the user’s electricity consumption portrait is realized to verify the effectiveness of the method.