复杂社交网络环境下基于社区演化和传递效应的推荐策略研究

项目名称： 复杂社交网络环境下基于社区演化和传递效应的推荐策略研究

项目编号： No.71471127

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 管理科学

项目作者： 田津

作者单位： 天津大学

项目金额： 60万元

中文摘要： 社交网络的发展影响着电子商务中网络营销模式的改变，社交网络环境下的个性化推荐研究具有重大的理论意义和应用价值。本项目拟针对复杂社交网络环境下的个性化推荐展开深入研究。首先，研究社交网络影响下用户兴趣和兴趣社区之间的关联关系，建立复杂社交网络环境下用户-社区协同演化模型，深入分析用户兴趣的生命周期，实现兴趣精准定位。其次，建立社交-用户-产品三层混合推荐模型，分析社交网络和产品的映射关系，扩展用户兴趣，挖掘产品的潜在用户群体。然后，研究社交网络和兴趣网络之间转化和影响，促进两个网络的融合，以提高社交网络影响下的推荐算法性能。最后研究两个网络中的推荐传递效应，建立基于社交网络和兴趣网络的混合推荐传递模型。本项目深入研究用户兴趣网络和社交网络中社区演化以及推荐传递效应，制定混合推荐策略，为复杂社交网络环境下的个性化推荐提供新思路，有助于电商企业全面分析和制定推荐策略，推动新的社交化商务模式产生。

中文关键词： 个性化推荐；社区演化；推荐传递效应；社交网络；电子商务

英文摘要： The social network has given a deeply influence on the E-commerce model. The research of personalized recommendation has been playing a significant role in both the theory and the application. This project aims to investigate the personalized recommendation in a social network context. Firstly of all, we will explore the relationship between the users' interest evolution patterns and the community evolution process, and develop a coevolution model of user and community based on the concept of the interest lifecycle. Moreover, it will discover mutual influence between the social network and the products, and propose a novel three-layer recommender model of social network, user and product. This model can mine both the potential user interest and the potential user group. The project will also study on the interaction between the social network and interest network in order to accelerate the two networks' merging. In addition, a transfer model will be developed by analyzing the transfer effect in and between the social network and the interest network. Finally, a hybrid recommendation model will be presented by fully considering the community evolution and the transfer effect, to predict the user profile and give the refine recommendation. It will be valuable for both the recommender systems and the social commerce.

英文关键词： Personalized recommendation;Community evolution;Transfer effect;Social network;E-commerce