面向C2C在线情境的一种个性化三维推荐方法

发布时间：2018-06-27 23:22

本文选题：三维推荐 + 个性化推荐　；参考：《情报学报》2016年06期

【摘要】：推荐系统是电子商务环境下过滤过载信息的有效工具,但只向用户推荐商品的传统二维推荐方法并不适用于C2C(Customer to Customer)电商情境。在C2C在线情境中,商品的供应者不止一个,消费者不但需要筛选商品,而且需要筛选卖家。针对这一需求,本文提出了一种新的个性化三维推荐方法,同时考虑C2C在线情境下买家、卖家和商品三者之间的相关性,并为买家提供卖家和商品的组合推荐。该方法包括四个阶段:首先基于卖家特征计算卖家相似度;其次依据卖家相似性和销售关系对三维推荐空间中的历史评分数据进行补充,降低其稀疏性;然后依据推导出的商品评分计算买家相似度,寻找具有相似商品偏好的最近邻;最后通过一个三维预测模型,计算买家对"卖家和商品"组合的未知评分,并根据预测评分进行推荐。通过基于淘宝网的真实数据实验,证实在C2C情境中本文提出的三维推荐方法比传统二维推荐方法更加有效。
[Abstract]:Recommendation system is an effective tool for filtering overload information in electronic commerce environment, but the traditional two-dimensional recommendation method which only recommends goods to users is not suitable for C2C (customer to customer) e-commerce situation. In a C2C online situation, there is more than one supplier of goods, and consumers need to screen not only the goods but also the sellers. In order to meet this demand, this paper proposes a new personalized 3D recommendation method, considering the correlation among buyer, seller and commodity in C2C online context, and provides the combination recommendation of seller and commodity for buyer. The method consists of four stages: firstly, the seller similarity is calculated based on the seller characteristics; secondly, the historical score data in 3D recommendation space are supplemented according to seller similarity and sales relationship to reduce its sparsity. Then the buyer similarity is calculated according to the derived commodity score, and the nearest neighbor with similar commodity preference is found. Finally, through a three-dimensional prediction model, the buyer's unknown score on the combination of "seller and commodity" is calculated. And according to the prediction score to recommend. Through the real data experiment based on Taobao, it is proved that the 3D recommendation method proposed in this paper is more effective than the traditional two-dimensional recommendation method in C2C context.
【作者单位】：广东工业大学管理学院;武汉大学信息管理学院;广东工业大学经济与贸易学院;
【基金】：国家社会科学基金青年项目“移动网络环境下情景敏感的个性化知识推荐机制研究”(11CTQ020) 广东省自然科学基金一般项目“移动泛在网络环境下基于情境语义的O2O实时推荐机制研究”(2015A030313499)
【分类号】：TP391.3

【参考文献】