Loading...
Search
| Friend's email | |
| Your name | |
| Your email | |
| enter code | |
This page was sent successfuly
Digital Object List
Bookmark
Review List- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 48067 (19)
- University: Sharif University of Technology
- Department: Computer Engineering
- Advisor(s): Izadi, Mohammad
- Abstract:
- Nowadays recommender systems are one of the most important parts of big websites. These systems help users to find their intended items among enormous amounts of data. Traditionally, recommender systems are designed and implemented using different methods such as content based, collaborative and demographic filtering. Each of these methods had some problems that lead to emergence of a new kind of recommender systems called hybrid recommender systems. This kind of recommender systems try to combine the other methods and make them better. In this thesis, we have selected some previous recommender systems and then, we have made a new system by combining and reforming them. The resulted recommender system not only can resolve the problems of their constituent parts but also it has better performance. In the suggested recommender system that its main goal is e-shop websites, we used items attributes to resolve sparsity problem, demographic information, costumer lifetime value and items attributes to resolve cold start problem. In addition, we recognize the changes of users’ interests by using time contextual information. Finally, we use MovieLens dataset to evaluate the suggested recommender system. The results confirm its better performance in comparison with the previous studied works
- Keywords:
- Context Information ; Recommender System ; Collaborative Filtering ; Content Base Filtering ; Demographic Filtering ; Items Categories ; Interests Change
-
محتواي کتاب
- view