Predictive analysis using social media integration techniques

Abstract

This thesis intends to model users interests from his activities on social media as well as the networks he is a part of. Besides that, the thesis also intends to re ect the change in ones behaviors over a period. It is observed that ones interests vary with time, thus inferring it only from his own activity is inconsistent. But, if a users net work is also traced for similar and correlated actions, then it is possible to infer ones consistent interest to a certain extent. This thesis introduces an approach that will consider a user and his networks activities, common tags, people and organizations from their social media posts which prepare a strong ground for the desired inference. Furthermore, to enrich inferred interests, the approach mentioned here also utilizes the semantic web using DBpedia ontology