FCA for Users' Social Circles Extension.

Nowadays, more and more users share real-time information and news on social networks and micro-blogging such as Flickr, BlogCatalog and Livejournal. With the increasing number of registered users in this kind of sites, find relevant and useful information becomes difficult. As the solution to this information glut problem, recommendation technologies emerged to help the exchanging of reliable sources of information, products or documents. One of the main mechanisms for organizing users online social networks is to categorize their friends into what we refer to as social circles. Currently, users identify and extend their social circles manually, or in a basic way by identifying friends sharing common activities. Neither approach is really satisfactory since they are time consuming and may function poorly as a user adds more friends. In this context, we develop a new algorithm called RecomFriend for users' social circles extension and friends recommendation based on a set of selected Communities Of Shared Interests (COSI) using Formal Concept Analysis (FCA) techniques. Carried out experiments over real-world datasets emphasize the relevance of our proposal and open many issues.