Consumer Psychological Effect Integrated Personalized Commodity Recommendation Model Based on the Fuzzy C‐Means Convolutional Neural Network
提出一种融合消费者从众、第三方和边际递减三种心理效应的个性化商品推荐模型，利用模糊C均值卷积神经网络处理数据稀疏性，并通过实证验证其有效性。
ABSTRACT The personalized recommendation model explores potential interests by analyzing consumers' historical behavior. Consumer psychology is an internal factor of consumer behavior; therefore, it is crucial to take into account its effect when improving the accuracy of recommendation. This paper proposes personalized commodity recommendation model based on consumer psychological effects. The quantitative approach rests on the fuzzy c‐means convolutional neural network (FCM‐CNN‐CPE). First, to alleviate data sparsity, we integrate the geographical location information into a fuzzy c‐means clustering algorithm to cluster consumers. Second, combining with the triple psychological effects of consumption (conformity effect, third‐person effect and marginal diminishing effect), the improved rating index and consumption probability measurement adjust the original rating information. Building on these, we propose a CNN‐based personalized commodity recommendation process, and subsequently validate the proposed method's effectiveness through empirical analysis. The research contributes to the theoretical development of recommendation models integrating consumer psychological effects and offer guidance for personalized commodity recommendation in practical applications.