Mode of access: World Wide Web.

System requirements: Adobe Acrobat Reader.

Part of: Synthesis digital library of engineering and computer science.

Includes bibliographical references (pages 177-191).

1. Introduction -- 1.1 Background -- 1.1.1 Social theories -- 1.1.2 Social tie analysis -- 1.1.3 Social influence analysis -- 1.1.4 User modeling and actions -- 1.1.5 Graphical models -- 1.2 Book outline --

2. Social tie analysis -- 2.1 Overview -- 2.2 Predicting missing links -- 2.2.1 Similarity metrics -- 2.2.2 Matrix factorization -- 2.3 Inferring social ties -- 2.3.1 Problem formulation -- 2.3.2 Unsupervised learning to infer social ties -- 2.3.3 Supervised learning to infer social ties -- 2.3.4 Actively learning to infer social ties -- 2.3.5 Inferring social ties across heterogeneous networks -- 2.4 Conclusions --

3. Social influence analysis -- 3.1 Overview -- 3.2 Mining topic-level social influence analysis -- 3.2.1 Topical affinity propagation -- 3.2.2 Dynamic social influence analysis -- 3.2.3 Model application -- 3.2.4 Experimental results -- 3.2.5 Summary -- 3.3 Mining topic-level influence from heterogeneous networks -- 3.3.1 The approach framework -- 3.3.2 Influence propagation and aggregation -- 3.3.3 Conservative and non-conservative propagation -- 3.3.4 User behavior prediction -- 3.3.5 Evaluation -- 3.3.6 Summary -- 3.4 Conclusions --

4. User behavior modeling and prediction -- 4.1 Overview -- 4.2 Approach framework for social action prediction -- 4.2.1 Model learning -- 4.3 Evaluation -- 4.3.1 Evaluation metrics -- 4.3.2 Prediction performance -- 4.3.3 Efficiency performance -- 4.3.4 Qualitative case study -- 4.4 Summary --

5. ArnetMiner: deep mining for academic social networks -- 5.1 Overview -- 5.2 Researcher profile extraction -- 5.2.1 A unified approach to profiling -- 5.2.2 Profile extraction performance -- 5.3 Name disambiguation -- 5.3.1 A unified probabilistic framework -- 5.3.2 Name disambiguation performance -- 5.4 Topic modeling -- 5.4.1 Our proposed topic models -- 5.5 Expertise search -- 5.5.1 Data sets and evaluation measures -- 5.5.2 Results -- 5.6 Academic social network mining -- 5.6.1 Mining advisor-advisee relationships -- 5.6.2 Measuring academic influence -- 5.6.3 Modeling researcher interests -- 5.7 Conclusions --

6. Research frontiers -- A. Resources -- Software -- Data sets -- Bibliography -- Authors' biographies.

Abstract freely available; full-text restricted to subscribers or individual document purchasers.

Compendex

INSPEC

Google scholar

Google book search

Online social networks have already become a bridge connecting our physical daily life with the (web-based) information space. This connection produces a huge volume of data, not only about the information itself, but also about user behavior. The ubiquity of the social Web and the wealth of social data offer us unprecedented opportunities for studying the interaction patterns among users so as to understand the dynamic mechanisms underlying different networks, something that was previously difficult to explore due to the lack of available data. In this book, we present the architecture of the research for social network mining, from a microscopic point of view. We focus on investigating several key issues in social networks. Specifically, we begin with analytics of social interactions between users. The first kinds of questions we try to answer are: What are the fundamental factors that form the different categories of social ties? How have reciprocal relationships been developed from parasocial relationships? How do connected users further form groups? Another theme addressed in this book is the study of social influence. Social influence occurs when one's opinions, emotions, or behaviors are affected by others, intentionally or unintentionally. Considerable research has been conducted to verify the existence of social influence in various networks. However, few literature studies address how to quantify the strength of influence between users from different aspects. In Chapter 4 and in [138], we have studied how to model and predict user behaviors. One fundamental problem is distinguishing the effects of different social factors such as social influence, homophily, and individual's characteristics. We introduce a probabilistic model to address this problem. Finally, we use an academic social network, ArnetMiner, as an example to demonstrate how we apply the introduced technologies for mining real social networks. In this system, we try to mine knowledge from both the informative (publication) network and the social (collaboration) network, and to understand the interaction mechanisms between the two networks. The system has been in operation since 2006 and has already attracted millions of users from more than 220 countries/regions.

Also available in print.

Title from PDF title page (viewed on May 20, 2015).