Advisory Committee Chair
Leon Jololian
Advisory Committee Members
Murat Tanik
Karthik Lingasubramanian
Document Type
Thesis
Date of Award
2019
Degree Name by School
Master of Science in Electrical Engineering (MSEE) School of Engineering
Abstract
Increasingly people of all ages consume news and entertainment through electronic media and social media [1]. As the internet in general and social media in particular are a recent phenomenon, the laws governing them and the technologies to monitor them are still evolving. There is a general consensus on the ubiquity and power of these media, hence the worry of how to handle these media. In this context, validating the online content becomes of paramount importance. Fighting fake content is not only relevant to news and current affairs but is very useful in other areas like technical content, legal content etc. as the corpus of content in every field is exponentially increasing. While manual detection of fake content is the most accurate method of detection in any field, its scalability is nearly impossible [2]. Leveraging the current tools in technology allows for scalability, however accuracy suffers. In this paper we have attempted to create a general content validation framework to help fight disinformation. By combining the manual approach with the use of technology, our framework has a method that is scalable and fairly accurate. Our framework consists of two step approach; step 1 is cross-referencing, and step 2 is computational fact-checking which consists of checking facts against known and trusted databases. The framework has the ability of intaking content and giving users a probabilistic score on the validity of the content being analyzed. As a validation step of the framework itself, this paper takes news articles as its specific case study, as data for such validation is readily available. The early results of our validation look very promising. At the end we will also discuss improvements needed and future work opportunities to make this framework even more robust.
Recommended Citation
Patel, Adilahmed, "Framework For Validation Of Different Media Types Using A Model Based On Consensus" (2019). All ETDs from UAB. 2678.
https://digitalcommons.library.uab.edu/etd-collection/2678
