PROJECT TITLE :

User Vitality Ranking and Prediction in Social Networking Services: a Dynamic Network Perspective - 2017

ABSTRACT:

Social networking services are prevalent at several on-line communities such as Twitter.com and Weibo.com, where countless users keep interacting with each other each day. One attention-grabbing and necessary downside in the social networking services is to rank users based on their vitality in a timely fashion. An correct ranking list of user vitality might benefit many parties in social network services like the ads suppliers and site operators. Although it is very promising to obtain a vitality-based ranking list of users, there are many technical challenges due to the large scale and dynamics of social networking knowledge. In this paper, we have a tendency to propose a distinctive perspective to attain this goal, that is quantifying user vitality by analyzing the dynamic interactions among users on social networks. Examples of social network embody however aren't restricted to social networks in microblog sites and academical collaboration networks. Intuitively, if a user has many interactions together with his friends at intervals a time period and most of his friends don't have many interactions with their friends simultaneously, it is very probably that this user has high vitality. Based mostly on this concept, we have a tendency to develop quantitative measurements for user vitality and propose our first algorithm for ranking users based mostly vitality. Additionally, we have a tendency to any consider the mutual influence between users while computing the vitality measurements and propose the second ranking algorithm, that computes user vitality in an iterative way. Alternative than user vitality ranking, we tend to conjointly introduce a vitality prediction problem, which is additionally of nice importance for several applications in social networking services. Along this line, we develop a customized prediction model to solve the vitality prediction downside. To judge the performance of our algorithms, we tend to collect 2 dynamic social network data sets. The experimental results with both knowledge sets clearly demonstrate the advantage of our ranking and prediction ways.