Research On Analysis Method Of Network User Preference

With the rapid development and wide use of the Internet, the amount of information on the Internet is having an exponential growth. How to get the required information from the mass of information has become a serious problem. Personalized recommendation service provides an effective way to solve this problem, and getting user preferences is the premise of recommendation. So researching the methods of analyzing the network users’ preferences is very meaningful for providing better information service.Firstly, this paper introduced the development and the status quo of the researching on the analysis methods of network user preferences and the recommendation technologies, and illustrated the significance of this paper. Secondly, this paper chose the micro-blog data as the research object, based on the researching of the micro-blog user data and the micro-blog community. Then, this paper used the API interfaces and the web crawler technologies to achieve the automated collection of the micro-blog data.This paper made a detailed analysis on the Sina micro-blog data, and extracted the information which can represent the preferences of micro-blog users. The information includes five aspects:personal information, micro-blog text, user relationship, user communication information, user impact. This paper provided some methods to make the preferences into quantization, and saved them in Xml files. Then, this paper gave a personalization recommendation algorithm based on the user preferences which were saved in the Xml files. The algorithm mainly works in two aspects:the one is the content-based, use cosine similarity algorithm to calculation the similarity between users, this method uses LDA to get improvement, and does the further screening by user impact; the other aspect is relation-based, it uses the PersonalRank algorithm to calculate the similarity, and adds user communication behavior weighting to improve the algorithm. After get the results from the two aspects, the algorithm uses the weighted hybridization to get the final result. This paper also made an experiment to prove the feasibility of the algorithm. At last, this paper designed a recommended system for users based on the data of Sina micro-blog.This work has been supported by the National Natural Science Foundation of China under Grant61172072,61271308, and Beijing Natural Science Foundation under Grant4112045, and the Research Fund for the Doctoral Program of Higher Education of China under Grant W11C100030, the Beijing Science and Technology Program under Grant Z121100000312024, and Beijing Municipal Commission of Education Discipline Construction and Graduate Construction Project.