Mixed Clustering Algorithm and Electricity Customer Classification

The principle of cluster analysis combined kernel methods with data description was researched firstly, based on this, the cluster analysis process combined kernel methods with data description and artificial immune system of real encoding was put forward. Then identify indexes of electricity customer classification ware confirmed. Finally, according to a power company some customer data, electricity customers ware classed, using the new cluster analysis method. Electricity customers ware divided into three categories: the first category includes 4 electricity customers,the second category includes 7 electricity customers, the third category includes 10 electricity customers. The results show that the new cluster analysis method can class samples and calculate center of mass, the calculation process is simple, and avoiding different categories when is different in kernel function, the case randomly determining initial immune body and increasing the computation times is avoided also, the precision results can be obtained.