Veiculação de publicidade em redes sociais utilizando perfis de usuários

Abstract

Social Networks are among the most used services on the Web. Every day, millions of users insert personal information on websites such as Orkut and Facebook. Such information is of great importance, because the user is talking about himself, representing an explicit personal data. In this dissertation, we propose a model for advertising in social networks, using information contained in the profiles of its users. For this model, we proposed a two step approach: first, an approach for identifying entities using Wikipedia articles as source to filter and expand the information contained in the profiles; and then use machine learning to reshape the ranking of recommended advertisements.