Loading...

Computational Modeling and Mining of Trust and Reputation in Social Networks Based on Soft Computing Approach

Rabiee Kenari, Yasaman | 2011

2300 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 42410 (52)
  4. University: Sharif University of Technology, International Campus, Kish Island
  5. Department: Science and Engineering
  6. Advisor(s): Azmi, Reza; Sedighi Moshkenani, Mohsen
  7. Abstract:
  8. Nowadays trust has become one of the main challenges of every online social network. The open and decentralized nature of Peer-to-Peer network makes it vulnerable to be exploited by malicious users to distribute tampered and harmful contents. We propose a dynamic computational trust model based on a neuro-fuzzy approach and implemented it on Gnutella network. The model is a novel application of control systems approach for evaluating trust and reputation in P2P networks. We utilize Local Linear Model Tree (LOLIMOT) to perform a nonlinear dynamic system identiffcation and thus to approximate trust. The performance of our reputation-based trust model is evaluated by simulation. We create a network emulating Gnutella behaviour using MATLAB and conduct several tests in order to inspect it. The behaviour of malicious peers is modeled; several scenarios of malicious attacks are created. A nonlinear dynamic behaviour is added by changing global trust values during system operation. Our network is also compared with another computational model of trust and reputation. Simulation results confirm the robustness, adaptation, and accuracy of our modeled network
  9. Keywords:
  10. Reputation ; Social Networks ; Malicious Attack ; Reputation-Based Trust ; Local Linear Model Tree ; Peer-to-Peer File Sharing

 Digital Object List

  • محتواي پايان نامه
  •   view