Loading...
| Friend's email | |
| Your name | |
| Your email | |
| enter code | |
This page was sent successfuly
- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 45567 (19)
- University: Sharif University of Technology
- Department: Computer Engineering
- Advisor(s): Jalili, Mahdi
- Abstract:
- Recommender systems are in the center of network science and becoming increasingly important in individual businesses for providing efficient personalized services and products to users. The focus of previous research in the field of recommendation systems was on improving the precision of the system through designing more accurate recommendation lists. Recently, the community has been paying attention to diversity and novelty of recommendation list as key characteristics of modern recommender systems. In many cases, novelty and precision do not go in the same direction and the accuracy-novelty dilemma is one of the challenging problems in recommender systems, which needs efforts in making a trade-off between them.
In this poject, we introduce novel filters to select candidate items for recommendation in order to provide novel and accurate recommendation to users. We also introduce a probabilistic structure to resolve the novelty-accuracy dilemma in recommender systems. The proposed filters to select candidate items are based on their popularity, which is defined as the number of votes they receive in a certain time interval. Wavelet transform is used for analyzing popularity time series and forecasting their trend in future time steps. We introduce two filtering algorithms based on the information extracted from analyzing popularity time series of the items: Popularity-based filtering and novelty and popularity based filtering algorithms. Theese filters can be applied as add-on to any recommendation algorithm. Experiments showed that the algorithms could significantly improve both the accuracy and effective novelty of the classical recommenders.
Also, we proposed a hybrid model with adjustable level of novelty and precision such that by tuning a single parameter, there levels could be controlled. The proposed recommendation model consists of two models; one for maximization of the accuracy and the other one for specification of the recommendation list to the tastes of users. We showed that the proposed model outperforms classical recommendation algorithms not only in terms of accuracy and novelty metrics, but also in terms of other important performance indicators such as coverage and diversity. Our proposed methods could be extensively applied in real commercial systems due to their low computational complexity and significant performance.
- Keywords:
- Information Extraction ; Recommender System ; Accuracy ; Collaborative Filtering ; Diversity Recommender System ; Markov Chain Based Recommender System
Digital Object List-
محتواي کتاب - view
Bookmark