Description

The usage and applications of social media have become pervasive. This has enabled an innovative paradigm to solve multimedia problems (e.g., recommendation and popularity prediction), which are otherwise hard to address purely by traditional approaches. In this paper, we investigate how to build a mutual connection among the disparate social media on the Internet, using which cross-domain media recommendation can be realized. We accomplish this goal through Social Transfer a novel cross-domain real-time transfer learning framework. While existing transfer learning methods do not address how to utilize the real time social streams, our proposed Social Transfer is able to effectively learn from social streams to help multimedia applications, assuming an intermediate topic space can be built across domains. In the existing system, Frameworks exist that connect directly-related item domains, such as a music album and tags on that album, or web pages and queries to them. However, these cannot be applied to indirectly-related item domains in social networks, such as tweets and user labels. In this project, the system proposed a novel Hybrid Random Walk (HRW) method, which incorporates such factors, to select transferable items in auxiliary domains, bridge cross-domain knowledge with the social domain, and accurately predict user-item links in a target domain. This method can be naturally applied to graph-based applications such as social networks, information networks, and biological networks. Extensive experimentation on a large real social dataset demonstrates that HRW produces significantly superior recommendations for web posts on social net-works.