Abstract | ||
---|---|---|
In order to improve the accuracy of recommendation algorithm in social network applications, a new recommendation method based on traditional collaborative filtering recommendation algorithm, which called Trust-based Collaborative Filtering, is proposed and verified in this paper. Firstly, we analyze users' behaviors and relationships in social network, and propose a trust calculation method based on Dijkstra's algorithm. Secondly, we integrate users' trust information into the collaborative filtering algorithm to recommend in social network. Finally, we choose Flixster dataset to validate the proposed model and use the Mean Absolute Error (MAE) as the evaluation metric. Experiment results show that Trust-based CF significantly improves the recommendation quality in social network. |
Year | DOI | Venue |
---|---|---|
2016 | 10.1109/CITS.2016.7546412 | 2016 International Conference on Computer, Information and Telecommunication Systems (CITS) |
Keywords | Field | DocType |
trust-based collaborative filtering algorithm,social network,recommendation algorithm,trust calculation method,Dijkstra algorithm,mean absolute error,MAE,evaluation metric | Recommender system,Data mining,Algorithm design,Collaborative filtering,Social network,Computer science,Mean absolute error,Algorithm,Prediction algorithms,Dijkstra's algorithm | Conference |
ISSN | ISBN | Citations |
2326-2338 | 978-1-5090-0691-5 | 1 |
PageRank | References | Authors |
0.40 | 7 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Xinxin Chen | 1 | 1 | 0.40 |
Yu Guo | 2 | 6 | 5.89 |
Yang Yang | 3 | 612 | 174.82 |
Zhenqiang Mi | 4 | 5 | 6.54 |