On enhancing recommender systems by utilizing general social networks combined with users goals and contextual awareness

Rana Chamsi Abu Quba 1, 2
2 SMA - Systèmes Multi-Agents
LIRIS - Laboratoire d'InfoRmatique en Image et Systèmes d'information
Abstract : We are surrounded by decisions to take, what book to read next? What film to watch this night and in the week-end? As the number of items became tremendous the use of recommendation systems became essential in daily life. At the same time social network become indispensable in people’s daily lives; people from different countries and age groups use them on a daily basis. While people are spending time on social networks, they are leaving valuable information about them attracting researchers’ attention. Recommendation is one domain that has been affected by the social networks widespread; the result is the social recommenders’ studies. However, in the literature we’ve found that most of the social recommenders were evaluated over Epinions, flixter and other type of domains based recommender social networks, which are composed of (users, items, ratings and relations). The proposed solutions can’t be extended directly to General Purpose Social Networks (GPSN) like Facebook and Twitter which are open social networks where users can do a variety of useful actions that can be useful for recommendation, but as they can’t rate items, these information are not possible to be used in recommender systems! Moreover, evaluations are based on the known metrics like MAE, and RMSE. This can’t guarantee the satisfaction of users, neither the good quality of recommendation
Complete list of metadatas

Cited literature [193 references]  Display  Hide  Download

Contributor : Abes Star <>
Submitted on : Tuesday, December 1, 2015 - 11:33:08 AM
Last modification on : Wednesday, November 20, 2019 - 3:03:52 AM
Long-term archiving on: Wednesday, March 2, 2016 - 12:11:43 PM


Version validated by the jury (STAR)


  • HAL Id : tel-01236089, version 1


Rana Chamsi Abu Quba. On enhancing recommender systems by utilizing general social networks combined with users goals and contextual awareness. Networking and Internet Architecture [cs.NI]. Université Claude Bernard - Lyon I, 2015. English. ⟨NNT : 2015LYO10061⟩. ⟨tel-01236089⟩



Record views


Files downloads