dc.contributor.author
Tsarapatsanis, Vaios
en
dc.date.accessioned
2020-06-04T12:40:29Z
dc.date.available
2020-06-05T00:00:29Z
dc.date.issued
2020-06-04
dc.identifier.uri
https://repository.ihu.edu.gr//xmlui/handle/11544/29500
dc.rights
Default License
dc.subject
Fake news detection
en
dc.subject
Ranking Model
en
dc.subject
Machine learning
en
dc.title
Fake news Detection
en
heal.type
masterThesis
en_US
heal.creatorID.email
tsarvakis@gmail.com
heal.generalDescription
This paper provides an experimental analysis with 5 famous well-known classifiers for the detection of fake news. The results are discussed and evaluated by appropriate metrics. The innovation of this study, the Ranking Model approach is capable of labeling new inputs as fake or real.
en
heal.license
http://creativecommons.org/licenses/by-nc/4.0
en_US
heal.references
Benjamin Riedel, I. A. (2018, May 21). A simple but tough-to-beat baseline for the Fake News Challenge stance detection task.
en_US
heal.recordProvider
School of Science and Technology, MSc in Data Science
en_US
heal.publicationDate
2018-12-10
heal.abstract
This dissertation was written as a part of the MSc in Data Science at the International Hellenic University. The study is based on fake news detection with machine learning concepts. Literature review on fake news was conducted in order to review the most significant theory concepts and realize the level of advancement regarding this topic by examining related work. A total number of 940 data points were extracted through a daily web scrapping procedure. The research part provides an experimental analysis with 5 well known classifiers and results are evaluated by appropriate metrics. Finally, the last part of the study is referring to the innovation of this study, the Ranking Model approach, which is capable of labeling new inputs as fake or real.
en
heal.advisorName
Tjortjis, Christos
en
heal.committeeMemberName
Berberidis, Christos
en
heal.committeeMemberName
Baltagiannis, Agamemnon
en
heal.academicPublisher
IHU
en
heal.academicPublisherID
ihu
en_US
heal.numberOfPages
81
en_US