Loading...

Enhancing Recommender Systems Using Analysis of Groups' Influence on Users in Social Networks

Nasr Esfahani, Hassan | 2015

485 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 47774 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Habibi, Jafar
  7. Abstract:
  8. Rapid growth in science and technology has created vast diversity in products, services and needs of people and groups. As the Internet and web technologies have progressed, a web-based solution to explore products of either a specific domain or multiple domains is mandatory. One of the main challenges in these systems is learning users’ preferences to recommend items possibly interesting to the user. Social network of the users is one of the sources that can inject additional information about them which can be exploited to improve accuracy of the system. Depending on the method, the performance, accuracy and personalization may differ .One of the most popular methods to extract this information from social networks is through community detection. These methods are both efficient and accurate enough, compared to the state-of-the-art methods. In this work we propose two different methods to enhance recommender systems by the help of community detection in social networks. In the model-based approach we exploit hierarchical community detection and in memory-based approach we propose a text retrieval based method to find users’ similarities with communities for recommendation which is both online and efficient while achieving adequate performance
  9. Keywords:
  10. Machine Learning ; Community Detection ; Information Retrieval ; Recommender System ; Community Detection ; Social Networks ; User Preferences

 Digital Object List

 Bookmark

...see more