Impact of Credibility on Opinion Analysis in Social Media

Issue title: Model and Data Engineering, MEDI 2016 Special Issue

Guest editors: Ladjel Bellatreche and Zbigniew W. Ras

Article type: Research Article

Authors: Ennaji, Fatima Zohra^{a; *} | Azaza, Lobna^b | Maamar, Zakaria^c | El Fazziki, Abdelaziz^d | Savonnet, Marinette^e | Sadgal, Mohamed^f | Leclercq, Eric^g | Amarouche, Idir Amine^h | Benslimane, Djamalⁱ

Affiliations: [a] LISI Lab, University of Marrakech, Morocco, [email protected] | [b] LE2I Lab, Dijon University, France, [email protected] | [c] Zayed University, Dubai, U.A.E, [email protected] | [d] LISI Lab, University of Marrakech, Morocco, [email protected] | [e] LE2I Lab, Dijon University, France, [email protected] | [f] LISI Lab, University of Marrakech, Morocco, [email protected] | [g] LE2I Lab, Dijon University, France, [email protected] | [h] USTHB, Algiers, Algeria, [email protected] | [i] LIRIS Lab, University Claude Bernard Lyon 1, France, [email protected]

Correspondence: [*] Address for correspondence: LISI Lab, University of Marrakech, Morocco

Abstract: In conjunction with the rapid growth and adoption of social media, people are more and more willing to share their personal experiences and opinions about products and/or services with the community. Opinions could be the basis of developing systems that would advise future users on how to proceed with any purchase without risking any disappointment. Unfortunately, opinions are not always genuine due to for instance, biased users as well as mixed feedback coming from the same users (i.e., multi-identity). This paper presents an approach for opinion analysis using credibility as a decisive criterion for supporting future users make sound decisions. The effectiveness of this approach has been tested using opinions posted on Twitter.

Keywords: Credibility, Multi-identity, Opinion, Twitter, Social media

DOI: 10.3233/FI-2018-1725

Journal: Fundamenta Informaticae, vol. 162, no. 4, pp. 259-281, 2018