Skip links

Emotional News Recommender System

Nowadays considering improvements in technology makes users have access to news through the internet, hence, news websites have achieved more popularity. However, According to the amount of data that are created each day, users are facing a huge amount of information updating continuously. Also because of the time that users must spend to find their right information, it makes a negative impact on performance. Thus, it could be helpful if an intelligent assistant helps users when they are looking for their desire data.

Recommender systems are software tools and techniques providing suggestions for items to be of use to a user. In this research, a new system has been introduced which recommends news based on news’ emotion in addition to users’ preferences and their past histories. It is aimed to make a positive impact on users and attempting to deliver their interested news. For this purpose, a hybrid recommender has been used that was a combination of content based approach and collaborative based method. In this system for predicting similarities between users’ preferences and recent news, TF-IDF and vector space model have been implemented. Also, a validation factor has been used to give more credit to data which represent users’ interest in a way that those data which are older than one month receives a validation factor equal to zero. This system has been evaluated in the term of performance and the effect it could have on users. According to these parameters, recall, precision and F1-score, which obtained from evaluating system’s performance, its F1-score average was %30.5 therefore, it seems that its performance is acceptable. Moreover, its effect on users’ emotion through a questionnaire has been analyzed. These analyze revealed that the system can have a positive impact on users’ emotion successfully.

Publication

  • Hakimi, A., Kazemifard, M., and Asghari, M., (2016). “EmoNews: an Emotional News Recommender System”. Journal of Digital Information Management, 14 (6), pp: 392-402.
  • Hakimi, A., and Kazemifard, M. (2015). “Emotional News Recommender System”. In Proceeding of the 6th International Conference of Cognitive Science (ICCS), Institute for Cognitive Science studies, Tehran, Iran, pp. 37-41

Join the Discussion

Return to top of page