Loading...

A Novel Context-Aware Model to Improve Quality of Recommender Systems

Abbasi, Ali | 2014

2056 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 46764 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Rabiei, Hamid Reza; Jalili, Mahdi
  7. Abstract:
  8. As the amount of data on the Internet grows, users face diverse options while searching for their desired information and items. Therefore, accessing what one is looking for, is usually time consuming and even impossible in some cases. In order to solve this issue, the goal of recommender systems is to offer recommendations which are compatible with users’ needs and preferences. One of the most important challenges of recommender systems is to improve the quality of recommendations. Recommender systems’ quality can be assessed using different metrics including precision, novelty and coverage. However, these metrics are inconsistent in some applications and improving one will cause a decline in others. The main contribution of this thesis is to provide a context-based model to increase coverage and novelty of recommendations while maintaining the precision of the system. Users’ mood, time and location are among the most used contextual information, whereas our primary focus is on location. The main concept used here is to integrate users’ individual behaviors in order to form group behaviors. This approach provides additional information which could not be extracted through analyzing users’ individual behavior. This information will improve the quality of recommendations. One can model users’ behavior using the list of accessed items. Therefore, grouping items and analyzing users’ behavior in group level leads to behavior integration. Spatial distance is used as the distance metric for items clustering. Moreover, for studying the integrated behavior in multi-resolutions, a hierarchical clustering method is utilized. Finally, we use a Bayesian approach to combine information from different levels for the sake of generating recommendation lists. Our simulation results indicate that the proposed method not only provides high accuracy, but also has high values of coverage and novelty considering other present methods
  9. Keywords:
  10. Machine Learning ; Recommender System ; Complex Network ; Spatial Information ; Context Information

 Digital Object List

 Bookmark

...see more