USER VITALITY RANKING AND PREDICTION IN SOCIAL NETWORKING SERVICES: A DYNAMIC NETWORK PERSPECTIVE
ABSTRACT
Social networking services have been prevalent at many online communities such as Twitter.com and Weibo.com, where millions of users keep interacting with each other every day. One interesting and important problem in the social networking services is to rank users based on their vitality in a timely fashion. An accurate ranking list of user vitality could benefit many parties in social network services such as the ads providers 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 data. In this paper, we propose a unique perspective to achieve this goal, which is quantifying user vitality by analyzing the dynamic interactions a users on social networks. Examples of social network include but are not limited to social networks in microblog sites and academical collaboration networks. Intuitively, if a user has many interactions with his friends within a time period and most of his friends do not have many interactions with their friends simultaneously, it is very likely that this user has high vitality. Based on this idea, we develop quantitative measurements for user vitality and propose our first algorithm for ranking users based vitality. Also we further consider the mutual influence between users while computing the vitality measurements and propose the second ranking algorithm, which computes user vitality in an iterative way. Other than user vitality ranking, we also introduce a vitality prediction problem, which is also of great importance for many applications in social networking services. Along this line, we develop a customized prediction model to solve the vitality prediction problem. To evaluate the performance of our algorithms, we collect two dynamic social network data sets. The experimental results with both data sets clearly demonstrate the advantage of our ranking and prediction methods.
EXISTING SYSTEM:
Related work can be grouped into two categories. The first category is most relevant that includes the work on measuring and ranking user in social network system. The second category is about the work on mearsuring user in network system. First, the user ranking algorithm in social network system has drawled a lot of attention in the research literature.The best known node ranking algorithms are Pagerank and HITS. Sergey Brin and Lawrence Page [2]proposed the pagerank to rank websites on the Internet. Pagerank is a link analysis algorithm which based on the directed graph(webgraph). The rank value indicates an importance of a particular node that represent the like-hood that users randomly clicking will arrive at any particular node. And, in , the authors presented two sampling algorithms for PageRank efficient approximation: Direct sampling and Adaptive sampling. Both methods sample the transition matrix and use the sample in PageRank computation. The hyper-link-induced topic search(HITS) was developed by Jon Kleinberg. This algorithm is a link analysis algorithm which rank the webpages. The authors presented a set of algorithms tools for rating and ranking the webpages from the directed graph of Internet environments. Furthermore, this work proposed a formulation of the notion of authority. PageRank/HITS is to find important websites that are linked to more different important websites and they do not consider the difference of nodes contribution to links at all, but in this paper we want to find those nodes that relatively contribute more to the interactions linked to them. However, . Meeyoung Cha et al. proposed a method to measure the user influence in Twitter used the directed links information, and present the comparison of three static measures of influence. However, them investigate the dynamics of user influence across topics and time which give a guide to the following research
PROPOSED SYSTEM:
in this paper, we propose two types of node vitality ranking algorithms that analyze the vitality of all nodes in a collective way. First, for a node A that has many interactions with his friends in a time period, if most of his friends do not have many interactions with their friends, it is very likely that the node A has high vitality. Based on this intuition, we define two measurements to quantify the vitality level of each node and propose the first algorithm. Second, by exploiting the mutual dependency of vitality a all users within a social network, we propose the second algorithm that infers the vitality level of users in an iterative way. Through the iteration, all nodes’ measurements propagate through the network and affect each other. Thus the second algorithm is able to collectively analyze the vitality score of all nodes by considering the whole network. Furthermore, upon our in-depth understanding about user vitality, we propose an improved model to predict the vitality of users. The successful prediction results will further benefit many applications on social networking sites. Finally, we conduct intensive ex- periments on both user vitality ranking and prediction with two large-scale real world data sets. The experimental results demonstrate the effectiveness and efficiency of our methods.
CONCLUDING REMARKS
In this paper, we presented a study on user vitality ranking and prediction in social networking services such as microblog application. Specifically, we first introduced a user vitality ranking problem, which is based on dynamic interactions between users on social networks. To solve this problem, we developed two algorithms to rank users based on vitality. While the first algorithm works based on the developed two user vitality measurements, the second algorithm further takes into account the mutual influence a users while computing the vitality measurements. Then we presented a user vitality prediction problem and introduced a regressionbased method for the prediction task. Intensive experiments on two real-world data sets that are collected from different domains clearly demonstrate the effectiveness of our ranking and prediction methods. The accurate results of both user vitality ranking and prediction could benefit many parties in different social networking services, e.g., a user vitality ranking list could help ads providers to better display their ads to active users and reach more audiences.
REFERENCES
[1] Eytan Bakshy, Jake M Hofman, Winter A Mason, and Duncan J Watts. Everyone’s an influencer: quantifying influence on twitter. In Proceedings of the fourth ACM international conference on Web search and data mining, pages 65–74. ACM, 2011.
[2] Sergey Brin and Lawrence Page. Reprint of: The anatomy of a largescale hypertextual web search engine. Computer networks, 56(18):3825– 3833, 2012.
[3] Robert Goodell Brown. Smoothing, forecasting and prediction of discrete time series. Courier Corporation, 2004.
[4] Christopher S Campbell, Paul P Maglio, Alex Cozzi, and Byron Dom. Expertise identification using email communications. In Proceedings of the twelfth international conference on Information and knowledge management, pages 528–531. ACM, 2003.
[5] Meeyoung Cha, Hamed Haddadi, Fabricio Benevenuto, and P Krishna Gummadi. Measuring user influence in twitter: The million follower fallacy. ICWSM, 10(10-17):30, 2010.
[6] Pedro Domingos and Matt Richardson. Mining the network value of customers. In Proceedings of the seventh ACM SIGKDD international conference on Knowledge discovery and data mining, pages 57–66. ACM, 2001.
[7] Philip E Brown Junlan Feng. Measuring user influence on twitter using modified k-shell decomposition. 2011.
[8] Jian Jiao, Jun Yan, Haibei Zhao, and Weiguo Fan. Expertrank: An expert user ranking algorithm in online communities. In New Trends in Information and Service Science, 2009. NISS’09. International Conference on, pages 674–679. IEEE, 2009.
[9] Jon M Kleinberg. Authoritative sources in a hyperlinked environment. Journal of the ACM (JACM), 46(5):604–632, 1999.
[10] Shamanth Kumar, Fred Morstatter, and Huan Liu. Twitter data analytics. Springer, 2014