Design And Implementation Of The Microblog User Tag Recommendation System Based On Similarity Analysis Of Attention

In recent years, with the more rapid development of science and technology in China, the microblog plays a very important role in our society and life. The microblog users can send and receive large amounts of information, and can also interact with other users, to achieve real-time information, commented, transfered and other operations. Due to the enormous amount of information generated by microblog every day, how to dig out the valuable information from a large number of microblog information, is a wide range industry in the modern Internet applications. In this industry, the value of user interest feature on the dig microblog is very high. However the traditional user interest mining method can not adapt to the characteristics of microblog, so based on the existing micro-blog interest mining method,the thesis designs a microblog users tag recommendation system which based on the similarity analysis of followers, according to the user’s label information to display the user’s interest.The mining of followers in this thesis, mainly includes the current user’s followers and the fans of the current user. The main contents of the thesis include:(1) According to the characteristics of microblog, the thesis designs a microblog users tag recommendation system which based on the similarity analysis of followers, firstly, the overall framework of the system is designed, then, on the microblog data acquisition mode, the thesis mainly uses the open platform API interface of the sina microblog, supplemented by web crawlers way to get user information.(2) The preprocess of the user’s data and information obtained, the repeated and useless information which without help to the recommendation result will be delete before is stored in the database. Then, using the SQL Server2007 database to store the acquired user information.(3) In label recommendation, the thesis mainly analyzes a label recommended algorithm which based on the combination of TF-IDF and weights,and analyzes the TagRan algorithm which make recommendation based on the interactions between users. Finally, combining the advantages of two algorithms, the thesis proposes the microblog users tag recommendation algorithm which based on the similarity analysis of followers, through experimental verification, the recommendation got by the proposed algorithm was more accurately than the above two algorithms.