SenFact algorithm: Fact-checking by the confrontation of opinions
Edouard Ngor Sarr, Ousmane Sall, Assane Diagne · 2017
The pluralism of media and the unbridled running in production have resulted in a considerable increase of the volume of treated and produced data, but also the proliferation of wrong information. The internet-user is so confronted with two major problems: the identification of the reliable sources and the reliability of information. Indeed, there is no guarantee of the accuracy of the given information and it is not always obvious to determine in advance priori one or several reliable sources. It has been since then necessary to have new methods, automated by fact-checking to satisfy this strong request with quality information. One of the approaches used in this quality search is Fact-Checking or «check on facts1». But, when the number of sources becomes numerous, most part of the algorithm of literature becomes less definite. In this paper, we introduce an algorithm of fact-checking named SenFact based on the confrontation of opinions that is to say the use of the knowledge of domain as a key element in the process of fact-checking.