Abstract | ||
---|---|---|
Recommending groups or communities to users can greatly improve the browsing experience in online photo sharing sites, e.g. Flickr. However, directly applying collaborative filtering techniques to group recommendation will suffer from "cold start" problem since many users do not affiliate to any groups. In this paper, we propose a hybrid recommendation method named Content-boosted Maximum Margin Matrix Factorization (CM3F), which combines collaborative user-group information with user similarity obtained from their uploaded images. Therefore, CM3F not only inherits the advantages of the state-of-the-art Maximum Margin Matrix Factorization (MMMF) method, but also owns the merits of the content-based graph regularization. The experiments conducted on our crawled dataset with 2196 users, 985 groups and 334467 images from Flickr demonstrate the effectiveness of our framework. |
Year | DOI | Venue |
---|---|---|
2013 | 10.1145/2513577.2513585 | DUBMOD@CIKM |
Keywords | Field | DocType |
browsing experience,group recommendation,cold start,content-boosted maximum margin matrix,maximum margin,content-based graph regularization,flickr group recommendation,collaborative user-group information,matrix factorization,state-of-the-art maximum margin,hybrid recommendation method,recommending group,social media | Data mining,Social media,Collaborative filtering,Information retrieval,Computer science,Upload,Matrix decomposition,Graph regularization,Cold start (automotive) | Conference |
Citations | PageRank | References |
0 | 0.34 | 5 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Yilun Wang | 1 | 297 | 13.03 |
Liang Chen | 2 | 313 | 36.77 |
Jian Wu | 3 | 933 | 95.62 |