Research on customer clustering for energy network marketing system based on K-means

Classifying customers based on customer information clustering plays is the most important resource for electronic marketing and online trading enterprises. Based on analyzing consumer characteristics and behaviors, a new algorithm for customer information clustering and customer classification is presented in the paper through improving the existing K-means algorithm. First a customer classification indicator system including 21 customer characteristics type indicators and customer behaviors type indicators is constructed; Second, the working principle and calculation flow of K-means algorithm are analyzed; Then some improvements such as algorithm calculation flow and initial cluster centers selection are advanced to overcome the limitations and defects of original K-means algorithm. Finally the paper realizes the improved K-means algorithm with the data from certain online trading enterprises and the experimental results verify that the improved algorithm can improve effectiveness and validity of customer information clustering and customer classification when used for customer management for online trading enterprises practically. © Sila Science.