面向C2C电子商务平台的三维个性化推荐方法研究

Research on Three-dimensional Personalized Recommendation Approach for C2C E-commerce Platform

定义C2C电子商务平台中不同于B2C平台的三维推荐空间和推荐问题,并针对该问题提出一种三维个性化推荐方法。该方法对传统二维协同过滤方法和基于内容推荐的方法进行混合和扩展。首先利用卖家特征属性计算卖家相似度,并基于销售关系和卖家相似度对三维评分数据集进行填补,以解决评分数据的稀疏问题,再利用填补后的评分数据计算买家相似度,获取最近邻并预测未知评分。实验证明,该方法能较好地解决C2C平台中的个性化推荐问题,在形成卖家和商品组合推荐时具有较好的性能。

This paper defines a three -dimensional recommendation space and recommendation task in C2C e -commerce platforms, which are different from B2C ones, and proposes a three - dimensional personalized recommendation approach, which extends the traditional two - dimensional collaborative filtering method and content - based recommendation method. The proposed approach firstly calculates seller similarities using seller features, and fills the three - dimensional rating set based on sales relations and seller similarities to solve the data sparsity problem. Then it calculates buyer similarities using historical ratings to decide neighbors and predict unknown ratings. A true data experiment proves that the proposed approach is effective to solve the personalized recommendation problem in C2C platforms and has good performanee when recommending seller and product combinations.