EgoTR: Personalized tweets recommendation approach

Abstract

Twitter and LinkedIn are two popular networks each in its territory. Nowadays, people use both of them in order to update their social (Twitter) and professional (LinkedIn) life. However, an information overload problem, caused by the data provided from these two networks separately, troubled many users. Indeed, the main goal of this work is to provide personalized recommendations that satisfy the user's expectations by exploiting the user generated content on Twitter and LinkedIn. We propose a method of recommending personalized tweet based on user's information from twitter and LinkedIn simultaneously. Our Final method considers two main elements: keywords extracted from Twitter and LinkedIn. Those extracted from Twitter are filtered by criteria such as hashtags, URL expansion and Tweets similarity. In order to evaluate our framework performance, we applied our system on a set of data collected from Twitter and LinkedIn. The experiments show that the proposed categorization of the elements is successfully important and our method outperforms several baseline methods.