隐式反馈场景中融合社交信息的上下文感知推荐

Implicit Feedback Personalized Recommendation Model Fusing Context-aware and Social Network Process

作为解决信息过载问题的有效方式,推荐系统能够根据用户偏好对海量信息进行过滤,为用户提供个性化的推荐。对如何利用隐式反馈数据进行个性化推荐进行了研究,提出了一种融合上下文信息和用户社交信息的隐式反馈推荐模型(Implicit Feedback Recommendation Model Fusing Context-aware and Social Network Process,IFCSP)。首先从数据集中提取与用户兴趣相关的上下文信息的属性集合,并以此作为分裂属性,使用决策树分类算法对"用户-产品-上下文"集合进行分类,从而将历史选择集合分组。对于要推荐的用户,根据其选择产品时的上下文信息,匹配最相似的分组,再使用基于隐式反馈的推荐模型(Implicit Feedback Recommendation Model,IFRM)预测用户对未选择产品的偏好,并结合用户的社交信息,进而对用户进行产品推荐。实验表明,该模型在平均正确率均值(MAP)和平均百分百排序(MPR)评价指标上均优于其他4种算法,可以显著提高系统的预测和推荐质量。

As a key solution to the problem of information overload,the recommender system can filter a large amount of information according to a user - s preference and provide personalized recommendations for users. This paper ex- plored the area of personalized recommendation based on implicit feedback and proposed a recommendation model, namely implicit feedback recommendation model fusing context-aware and social network process （IFCSP）, which is a novel context-aware recommender system incorporating processed social network information. This model handles contextual information by applying a decision tree algorithm to classify the original user-item-context selections so that the selections with similar contexts are grouped. Then implicit feedback recommendation model （IFRM） was employed to predict the preference of a user for a non-selected item using the partitioned matrix. In order to incorporate social net- work information,a regularization term was introduced to the IFRM objective function to infer a user＇s preference for an item by learning opinions from his/her friends who are expected to share similar tastes. The study provides comparative experimental results based on the typical Douban and MovieLens-1M data sets. Finally, the results show that the proposed approach outperforms state-of-the-art recommendation algorithms in terms of mean average precision （MAP） and mean percentage ranking （MPR）.