Loading...
Please wait, while we are loading the content...
Similar Documents
Collaborative Filtering Recommendation with Fuzzy-weighted User Similarity
| Content Provider | Semantic Scholar |
|---|---|
| Author | Lee, Soojung |
| Copyright Year | 2018 |
| Abstract | Similarity computation plays a critical role in collaborative filtering-based recommender systems. As these systems recommend items based on user ratings, they involve several inherent problems such as data sparsity, cold-start, scalability, and user subjectivity. Much effort has been devoted to handle these problems and simultaneously enhance the performance of the system, but there are still much to be improved. This study focuses on user-based collaborative filtering systems and proposes a new similarity measure which not only considers user ratings for common items but also reflects the rating behavior of all the users on each common item onto similarity. Performance of the proposed measure is investigated extensively under very different ratings data conditions. The results state that it mostly outperforms state-of-the-art similarity measures, where the degree of improvement is significantly high when it incorporates Pearson correlation. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://www.naun.org/main/NAUN/fuzzy/2018/a142017-060.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
