Research on E-commerce Customer Segmentation Based on the K-means++ algorithm

Abstract

In a long-term study, the author found that the attention and loyalty of e-commerce customers are very important factors for e-commerce customers to maintain and maintain. Therefore, on the customer value matrix AF of the traditional customer segmentation model, which represents the existing value, This article has added 2 variables that represent the value-added-potential of e-commerce customers, namely, the total clicks C representing the customer attention and the customer hold time H representing the customer loyalty, and constructs the AFCH e-commerce customer segmentation. The AFCH customer segmentation model is tested by K-means, SOM + K-means and K-means++ respectively. The error square and SSE were used as the algorithm's accurate measurement standard. The experimental effect found that the clustering results of the three algorithms were similar, The accuracy of K-Means++ algorithm is higher than that of the other two. Finally, this paper gives the experimental results of AFCH customer segmentation for an e-commerce enterprise by K-Means++ algorithm, which provides decision support for the customer management and specific marketing measures of e-commerce enterprises.