推荐系统是解决当今互联网信息过载问题的有效技术手段，具有重要的研究价值和广阔的应用前景。作为推荐系统领域的一个重要分支，情景感知的推荐系统（CARS）能够提供更准确的推荐服务，代表了推荐系统未来的主要发展方向。然而，CARS的研究仍处于起步阶段，面临着很多困难和挑战。本项目拟结合社交网络这一热点研究领域，对情景感知推荐技术进行较为系统化、理论化的探索，在融入社交信息的基础上，为CARS中的关键技术提供新的设计思路和解决方案。本项目拟重点研究"相关情景的界定"、"情景信息的获取"、"情景相关用户偏好的提取"、"情景感知的推荐生成算法"以及"CARS的效用评价"等几个方面，所提出的新方法，可望丰富CARS领域的理论成果，并为CARS在社交网络领域中的应用提供技术支撑。

Recommender system is an effective technique to solve the information overload problem on the internet, and thus has significant research value and broad application prospects. As an important branch of recommender system, the context-aware recommender system(CARS) represents the future development direction of recommender systems as it provides more accurate recommendations. However, research on CARS is still in its infancy, and is faced with many problems and challenges. This project intends to explore new context-aware recommendation techniques both systematically and theoretically considering the social network area, and will offer new ideas and solutions to critical technologies in CARS with the integration of social information. We mainly focus on five aspects, namely "The identification of related context", "Contextual information extraction", "The contextual user preferences extraction", "Context-aware recommendation generation algorithms", and "The evaluation of CARS". The proposed methods are expected on the one hand, to enrich the theoretical achievement in the field of CARS, and on the other to provide technical support for the applications of CARS in social networks.