USER PROFILE IN PERSONALIZING INFORMATION

In order to meet the users expectations, the information retrieval system classified the users following their profile then affect them to the appropriate chain access of information. Our approach is founded on the enrichment of users profiles that will be conducted from their social network which represent a rich source of information. However, one of our biggest challenge is when the user profile is newest or not very active, so the user profile may not contain all the interests and information that may be useful for a given mechanism. To bypass these challenges we developed a new mechanism that will allow us to detect similar users to this one and analyze their interest using a similarity techniques then algorithms that will improve its performance in a remarkable way by expanding user’s egocentric network which allow adding more nodes from different social networks to cover the entire user’s interest. Therefore, our proposal is to combine friends relationships of the user on different types of social network in a single egocentric social graph before the application of the algorithms to get a more complete user profile that will gather the majority of user’s interests.