基于关联规则与RFM聚类的社交电商用户购买行为分析

Analysis of Social E-Commerce Users’ Purchasing Behavior Based on Association Rules and RFM Clustering

摘要: 本文基于公开数据集网站天池提供的社交电商平台的真实数据，从用户特征、内容特征、社交特征及行为序列特征等多个维度对用户行为进行系统分析。首先，采用Apriori关联规则算法挖掘用户关键行为之间的关联关系，结果表明加购行为是用户购买决策中强烈的前置信号。然后基于RFM模型并结合K-means聚类算法对用户进行客户价值分层，在肘部法与轮廓系数法的支持下将客户划分为四类，并进一步通过AHP层次分析法与熵权法对客户价值进行综合评估。最后，在客户价值分层基础上，对不同用户群体的购买行为特征进行对比分析，发现价格与折扣因素对各类客户均具有普遍的正向影响，而视频内容与社交互动特征在不同价值客户群体中的作用存在差异。研究结果为社交电商平台实施精细化客户运营与差异化营销策略提供了数据支持与实践参考。

Abstract: This study is based on real-world data from a social e-commerce platform provided by the public dataset website Tianchi. User behavior is systematically analyzed from multiple dimensions, including user characteristics, content characteristics, social characteristics, and behavioral sequence characteristics. First, the Apriori association rule algorithm is applied to mine the relationships among key user behaviors. The results indicate that the add-to-cart behavior serves as a strong antecedent signal in users’ purchase decision-making process. Then, a customer value segmentation is performed using the RFM model combined with the K-means clustering algorithm. Supported by the elbow method and silhouette coefficient analysis, users are divided into four customer segments, and their customer value is further comprehensively evaluated using the Analytic Hierarchy Process (AHP) and the entropy weight method. Finally, based on customer value segmentation, comparative analyses of purchasing behavior across different user groups are conducted. The results show that price and discount factors have a universal positive impact on purchase behavior across all customer segments, while the effects of video content and social interaction features vary among customers with different value levels. The findings provide data-driven support and practical insights for implementing refined customer management and differentiated marketing strategies on social e-commerce platforms.