Extraction and recommendation of experts on topics of interest in social networks as an educational tool

© 2018 IEEE. Nowadays, Information and Communication Technologies (ICTs), along with Web 2.0 technologies, are enabling the globalization of Internet, providing a mean of access for the creation, dissemination and discussion of content. In this way, anyone can consult information of interest, assimilate it and turn it into useful knowledge. However, in recent years this new technological trend has driven users to generate a large amount of personal content, leaving aside any quality index, which translates into a new social problem known as information overload, infoxication or infobesity. The contact with information that is constantly increasing and of which validity has not been proven can cause difficulties, from the assimilation of knowledge to psychological disorders (anguish). The educational field is no stranger to this situation, as students use technology to support their academic processes. This research proposes the development of an experts recommendation tool (individuals who significantly manage a topic of interest) based on Twitter and Mendeley with a semi-supervised approach. In a Web application, keywords related to a topic of interest are entered and extracted from potential Mendeley experts, and then their accounts are located on Twitter. With this information, a user validates whether the Twitter profiles correspond to experts and authorizes the publication of a recommendation to students. With the semi-supervised approach, the accuracy of the recommendations is 100%, so the results obtained are promising.