Social Influence Analysis in Social Netw orking Big Data: - PowerPoint PPT Presentation

Social Influence Analysis in Social Netw orking Big Data: Opportunities and Challenges Presenter: Sancheng Peng Zhaoqing University

Contents Introduction 1 Relationship between SIA and BD 2 Framework of Social Influence Analysis 3 Properties of Social Influence 4 Opportunities 3 5 6 4 Existing work Challenges & Future Work 7

Introduction Social network: e.g. Online social network, mobile social network; Online social network: e.g. Facebook, Twitter, and LinkedIn; Mobile social network: e.g. Smartphone Social networking big data is a collection of very huge data sets with a great diversity of types from social networks (e.g., Facebook, call data).

Introduction 5V: Volume, Velocity, Variety, Value, and Veracity; Big data applications: physics, astronomy, atmospheric science, medicine, genomics, biology, biogeochemistry, and so on. Social influence: refers to the case when individuals change their behaviors under the influence of others.

Introduction The strength of social influence depends on the relation among individuals, network distances, time effect, characteristics of networks and individuals, etc. Social influence analysis is how to quantify the influence of each user and how to identify the most influential users in social networks. Applications: Viral marketing, online advertising, recommendation, expert finding, and so on.

Introduction Questions for Social influence analysis: 1) Who can be influenced?; 2) Who can influence whom?; 3) Who are vulnerable to be influenced? 4) Why is a user attracted to a particular group?; 5) Who are the most influential users in a specific social network?’’

Introduction Main idea of social influence analysis: 1) How to quantify the influence of each user; 2) How to identify the most influential k users; 3) How to put social influence analysis into actual application. …….

Introduction Important significance: 1) Helpful to understand the ways in which information, experiences, ideas, and innovations propagate across social networks. 2) Helpful to provide new insights into how people interact with and influence each other, and why their ideas and opinions on different subjects can spread in social networks.

Introduction 3) Helpful to understand social behaviors of people from the angle of sociology; 4) Helpful to provide a theoretical basis for making public decision and public opinion guidance; 5) Helpful to promote communication and dissemination of political, economic and cultural activities, as well as in other fields.

Relationship between SIA and BD

Framework of Social Influence Analysis Data collection from social networks Selection of evaluation Data preprocessing metrics Modeling and computing social influence Selection of the most influential top- k nodes Performance analysis: influential range, computation complexity

Properties of Social Influence 1) Dynamic It changes with the change of the context and time dynamically. 2) Propagative “word-of-mouth” propagation 3) Composable when several chains influence a member indirectly, then the influencer needs to compose the influence information. 4) Measurable referred to as the influence value or be represented with uncertainty

Properties of Social Influence 5) Subjective the biases and preferences of the influencer 6) Asymmetric if Bob influences Alice, it does not imply that Alice also influences Bob. 7) Event sensitive Influence may take a long time to build, but a single high-impact event may destroy it completely in a short time.

Opportunities for Social Influence Analysis 1) Promoting the development of social networks 2) Laying a sound theoretical foundation for research on social influence analysis 3) Facilitating collection and processing of social networking big data 4) Improving the application prospect of social influence analysis

Opportunities for Social Influence Analysis OSN Site Active Monthly Users Reported Date Facebook 1.55 Billion November, 2015 Youtube 1 Billion March, 2013 Google+ 540 Million October, 2013 Instagram 400 Million September, 2015 Twitter 320 Million September, 2015 Vine 200 Million September, 2015 Linkedin 187 Million April, 2014 Pinterest 100 Million September, 2015

Existing work for Social Influence Analysis 1) topic-oblivious social influence is measured either via the relative authority of individuals in their social network, or via the degree of information diffusion with the social network. 2) topic-based social influence is measured by counting how much information related to a topic may be propagated in the network. 3) pairwise-based social influence is defined based on social ties and interactions between users.

Existing work for Social Influence Analysis Selection of the most influential top-k nodes ----> Influence maximization problem ---->NP hard 1) Greedy algorithm 2) Heuristic algorithm 3) Extended algorithm

Existing work for Social Influence Analysis Influence spread 1) Cascade model 2) Threshold model 3) Scalable algorithm 4) Epidemic model 5) Voter model 6) Game-theoretic model

Existing work for Social Influence Analysis Take “Cascade model” as an example:

Challenges for Social Influence Analysis 1) Determining a set of effective evaluation metrics 2) Considering the characteristics of dynamic evolution in large-scale social networks 3) Characterizing casual relationship in large-scale social networks 4) Distinguishing positive influence, negative influence, and controversy influence

Challenges for Social Influence Analysis 5) Guaranteeing the efficiency and scalability on social influence analysis 6) Evaluating the influence of heterogeneous social network 7) Considering the collection and processing of social networking big data 8) Providing an effective mechanism to perform social influence analysis

撰写的论文 1. Smartphone Malware and Its Propagation Modeling: A Survey. IEEE Communications Surveys and Tutorials, 2014, 16(2): 925-941. 2. Propagation Model of Smartphone Worms Based on Semi-Markov Process and Social Relationship Graph. Computers & Security, 2014, Vol. 44, pp. 92-103. 3. Containing worm propagation with influence maximization algorithm in smartphones. Computer Networks, 2014, Vol. 74, pp. 103-113. 4. Mining Mechanism of Top- k Influential Nodes Based on Voting Algorithm in Mobile Social Networks. The 11th IEEE/IFIP International Conference on Embedded and Ubiquitous Computing (EUC 2013), November 13-15, 2013, pp. 2194-2199.

撰写的论文 5. Social Influence Analysis in Social Networking Big Data: Opportunities and Challenges. IEEE Network Magazine, 2015. (Accepted) 6. Entropy-Based Social Influence Evaluation in Mobile Social Networks. The 15th International Conference on Algorithms and Architectures for Parallel Processing (ICA3PP 2015), November 18-20, 2015, pp. 637-647. 7. Social influence modeling using information theory in mobile social networks. Submitted to Information Sciences, 2015. 8. Evaluation Modeling and Propagation Analysis on Social Influence Using Social Networking Big Data. Will be submitted to IEEE Transactions Dependable and Secure Computing, 2015.

Special Issue Topics: 1. Social influence evaluation in large-scale social networks 2. Social influence analysis in big data 3. Influence propagation in large-scale social networks 4. Dynamic social influence in large-scale social networks 5. Influence maximization problem with big data 6. User behavior analysis with social influence evaluation

Special Issue 7. Social influence analysis in heterogeneous social network 8. Casual relationship in large-scale social networks 9. Mechanism for distinguishing the positive, negative, and controversy influence 10. Models, methods, and tools for influence propagation 11. Malicious information propagation with social influence analysis 12. Secure social networking application with social influence analysis

Special Issue Title1: Social network analysis and its applications Title2: Social influence analysis and its applications ……

? Tel. +86-13727251176 ， E-mail: psc346@aliyun.com