Mapping different online behaviors to physical user for comprehensive knowledge-pushing services

With the quick development of the internet, it benefits a lot to our daily lives from the fast growing of applications. But on the other hand, time and energy of a specific user are limited and it is very difficult for him to find the interesting information from the massive data. In this paper, we focus on how to analyze the user's combined behavior characteristics from numbers of applications for obtaining his insightful interests and provide high quality of knowledge pushing services. Firstly, we employ the passive and the active measurement methods to extract the users' IDs of different applications. Secondly, we develop an ID association method named as “snow ball rolling” to mine the IDs belong to the specific physical user. Based on the mining results, we can combine the user's behavior in different applications together, which lay a solid foundation for insightful interest mining and efficiency knowledge pushing. Finally, experiment results based on datasets collected from the actual network show the efficiency and accuracy of our methods.