A framework for social networks new relationships recommendation from the use of folksonomies

This paper propose a new approach to approximate people with compatible interest and knowledge through the recommendation of new relations over a social network. The center idea of this paper is use folksonomy to describe the people profile, and recommendation based on collaborative filtering to find people with similar profile that current not being direct related in the social network. A framework was designed and a prototype developed. Experiments conducted with synthetic data and real users show how promising is the use of recommendation and folksonomy' techniques in the context of virtual social networks.