Loading...

Model Selection for Social Network Simulation in a Decision Support System

Aliakbary, Sadegh | 2015

1308 Viewed
  1. Type of Document: Ph.D. Dissertation
  2. Language: Farsi
  3. Document No: 46987 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Habibi, Jafar; Movaghar Rahimabadi, Ali
  7. Abstract:
  8. A social network represents a set of entities and their relationships. Telecommunication networks, online social networks, and paper citation networks are some examples of networks in real world. Nowadays, analysis of social networks is an interesting research area with important applications. Particularly, managers of the social networks and the decision makers often require intelligent decision support for futures study in these social systems. The demanded decision support systems make it possible to define the desired social problem and to analyze the ”what-if scenarios.“ Computer simulation is an appropriate approach toward such decision support systems. In this approach, the desired network is simulated, different scenarios are examined, and the results are studied. Simulation is an effective method for analysis of theories and to propose new hypotheses for social systems. The aim of this dissertation is to investigate open research problems towards a decision support system based on simulation of social networks. In the desired system, the decision maker can define the properties of the target network, synthesize an artificial network with the desired size and properties, run different social scenarios in the synthesized networks, and study the results of the simulations. Generation of customizable artificial networks is one of the main research problems in this dissertation. Artificial networks are stochastic graphs which are structurally similar to the target network. In this dissertation, we proposed a framework for generation of social networks based on some research steps. First, we proposed a novel method for feature extraction from the degree distribution of the social networks. Then, we proposed a similarity metric for comparing structure of social networks. The proposed similarity metric is utilized in our next step of research which involves generative model selection. Model selection is the art of finding the best generative model, among the candidates, which is capable of generating networks similar to the target network. The proposed model fitting method not only selects the best fitting model, but also tunes the model parameters in order to generate the best fitting network with the desired size. The proposed model fitting method is robust to noise. All of the proposed methods, which are based on machine learning, outperform the existing baselines with respect to accuracy and efficiency. In this dissertation, we have also presented an initial software reference model for scalable simulation of large social networks
  9. Keywords:
  10. Social Networks ; Data Mining ; Complex Network ; Multiagent System ; Generating Model ; Decision Making Support System ; Social Simulation

 Digital Object List

 Bookmark

...see more