Research Of Context-aware Recommendation Based On Rough Set Theory And Collaborative Filtering

With the development of Mobile Internet, Internet of Things, Ubiquitous Computing and E-commerce, the boundary of Internet is greatly expanded, and we are entering an age of ‘big data'. How to get the resources that mostly fit one's needs from among the vast amount of data, this is a very prominent problem. Existing traditional search engines and portal sites have alleviated this problem to some extent, but they still can not fully meet the requirements of all people. Recommender system is considered as an effective solution to solve the problem of "information overload" caused by massive data, it has got extensive attention and widely used of academia and industry, and many research results have been obtained.Context is an important factor in recommender systems. By introducing context information into the recommender system, Context-Aware Recommender System has both advantages of ‘ubiquitous computing' and ‘personalization'.It can further improve the recommendation accuracy and user satisfaction, and has important research significance and practical value. In existing CARS, the influence of different context attributes to the recommender system is ususally regarded as equal weight, however, the influence of different context type(like location or time) may be different. How to distinguish the influence of various contexts and integrate them into the recommending process to improve the prediction accuracy and recommendation quality of the recommender system, it is a very important research topic.In order to solve this problem, the main work that this article has done is as follows:(1) Firstly, this article has reviewed Context-Aware Recommendation technique, compared the similarities and differences between it and traditional recommendation techniques, and analyzed the whole process of Context-Aware Recommendation.(2)Secondly, in this paper, we proposed a context dimension reduction algorithm that based on rough set theory. By computing the significance of the context attributes, this algorithm can reduct insignificant attributes, thus increase the accuracy and efficiency of recommendation.(3)Thirdly, by integrating context information into collaborative filtering recommendation algorithm, this article designed a Rough Set based Context-Aware Recommendation algorithm. We constructed a user similarity calculation method that has included the context information, and realized the Context-Aware Collaborative Filtering Recommendation based on the user similarity.(4)Lastly, we compared our Rough Set based Context-Aware Recommendation algorithm with existing Context-Aware Recommendation algorithm, and conducted the simulation experiments. The experimental results demonstrate that our Context-Aware Recommendation algorithm has a significant effect on increasing the recommendation precision, and it's an effective algorithm.