Towards context-aware media recommendation based on social tagging

These days, there are huge amount of multimedia contents that are available for users. Selecting and exploiting favorite contents from such online available collections is a big challenge for users as well as smart engines due to the overwhelming huge number of available contents. This paper proposes a new model to help recommendation systems to select the most appropriate, favorite, or related contents by integrating contextual parameters. The proposed new context-aware recommendation model enriches the recommendation process with context to tailor the recommendation results to individual users. By leveraging the social tagging, our proposed model computes the latent preference of users on contexts from other similar contexts, as well as latent assignment of contexts for items from other similar items. By finding the similarities between the user's contexts and among the contexts and items, we can determine the attractive items given a particular context. We then map the context on the items depending on that particular user, in order to recommend the most relevant item suitable to the user's needs. Our evaluation results have shown a potential improvement to the recommendation quality compared to state-of-the-art recommendation algorithms that consider contextual information. We also compare the proposed model to other approaches where user's context is not used for personalizing the recommendation results.