Time-Aware User Profile Enrichment in the Collective Intelligence Context

The approaches of user profile enrichment come up with novel algorithms to better understand users' preferences. It allows business advertisement companies collecting collective intelligence data resulting from crowdsourcing. The enrichment is based on adding tags to the profiles. However, it stills unable to better understand the heterogeneity of users' interests. The exploration of the proximity between resource topics and user's interests will improve the profile enrichment methods. In this paper, we present a tag-based profile enrichment approach with the addition of time score describing the long and short term criteria of interest. We use graph analytics to generate graphs of users using the Louvain method by inspecting similar tags with a high time score. We created sub-nets of users with the same relevant interests. In future works, relevant semantic profiles description aims to match contents and users in order to make a relevant recommendation. Our approach helps companies to make their conclusions. The datasets Movielens and the dataset Mendeley have been conducted to evaluate the effectiveness of our approach.