PROJECT TITLE :

Collaborative Filtering-Based Recommendation of Online Social Voting - 2017

ABSTRACT:

Social voting is an emerging new feature in on-line social networks. It poses distinctive challenges and opportunities for recommendation. During this paper, we tend to develop a set of matrix-factorization (MF) and nearest-neighbor (NN)-primarily based recommender systems (RSs) that explore user social network and cluster affiliation information for social voting recommendation. Through experiments with real social voting traces, we tend to demonstrate that social network and cluster affiliation information will significantly improve the accuracy of popularity-primarily based voting recommendation, and social network data dominates cluster affiliation data in NN-based approaches. We have a tendency to also observe that social and cluster data is much more valuable to cold users than to heavy users. In our experiments, straightforward metapath-primarily based NN models outperform computation-intensive MF models in hot-voting recommendation, while users' interests for nonhot votings can be higher mined by MF models. We have a tendency to further propose a hybrid RS, bagging different single approaches to attain the most effective prime-k hit rate.