Abstract:

With the arrival of the era of Web 2.0,the popularity of microblog brings many opportunities and challenges to the public security work on the actual combat.In order to control the data of microblog user relationship effectively,we can use it on lots of public security business such as security early warning, prevention of crime and case detection.First of all,this paper defined the position and significance of the system of microblog user relationship analysis based on interest macroscopically, and designed the function and hierarchical structure of this system.Then,this paper studied two central technological modules in this system. Firstly taking the microblog user’s interest as a starting point, we divided microblog user into different communities by the K-means clustering algorithm (including preprocessing by PCA descending dimension algorithm) and diagram algorithm. Secondly,we analyzed the importance of microblog user in the communities by the important node analysis technology. In this paper,the system can divide microblog user into different communities theoretically according to the needs,and it can get the key users of each communities,which will provide a lot of convenience and support for the public security work.

Key words: microblog, interest, user relationship, prototype system