Hybridizing Social Filtering for Recommender Systems

Publisher:
Springer Berlin Heidelberg
Publication Type:
Conference Proceeding
Citation:
Foundations of Intelligent Systems, 2014
Issue Date:
2014
Full metadata record
Files in This Item:
Filename Description Size
Thumbnailpaper.pdfAccepted Manuscript version494.11 kB
Adobe PDF
Users send requests to recommender systems for getting suggested products or services. Collaborative filtering is a popular technique for making such suggestions efficiently, but it suffers from a drawback known as “cold-start” problem. Social filtering may succeed for such users, since it utilize the extra social relations of users. It gives us opportunities to eliminate the limitations by hybridizing social filtering into traditional collaborative filtering. To handle this issue, differing from previous fusion models that only combine the final results, this paper proposed a new neighborhood fusion model to make hybridization at an earlier and deeper stage. Experiment-based comparative analyses are also conducted. The results show that our model is of a higher recommendation quality, on different datasets.
Please use this identifier to cite or link to this item: