Modeling And Pattern Analysis On User Behaviors In Mobile Device And Social Network Enviroment

With the rapid development of computer technology, a variety of internet service providers open more and more flow inlet in the different form of hardware and software, like PC, smart tablet, mobile phones, electronic commerce, social network, game.etc. that means in now days Internet environment, there are many channels user may leave their tracks, in some degree, which could reflect the user’s personalization features about the online behaviors. Mining these personalization features may help us to understand the user’s daily habits better, and with the mining result, the work of service recommended could be more accurate.Users online behavior of using service generally occurs in two scenarios, One is about the individual user behavior just produced to satisfy their own needs, which there is not direct relations with others, such as online music, mobile reading. Another is that multiple users collaborate with each other to meet some certain requirements, such as social networks. these two scenarios named user behavior in mobile device and social network are studied through modeling and behavior pattern analysis.For researching the user behavior in mobile device, we select the android users as our target group and doing the following work: we collect the app usage with the context information; Proposed four kinds of patterns based on the context of user behaviors; Designing the mining algorithm based on behavior patterns and context, then evaluating the algorithm; proposing number of prediction strategies about behavior in mobile device; displaying the mining results through visualization system.For researching the user behavior in social network, we choose github which is the representative social collaboration platform with more behavior type and higher user activity to study and carried out the following work: collecting user behavior data through Crawling information technology; analyzing the characteristics of online users behavior in social network; Propose two behavior patterns: the participant-oriented patterns(PP) and role-oriented behavior patterns(RP); Designing the mining algorithm based on DP and AP, then evaluating the algorithm; proposing number of prediction strategies about user behavior in social network; displaying the mining results through visualization system.The results found that users’ online behavior follow some of patterns, the patterns can be better explained combined with context, and users collaboration in social network both have difference and similarity.