Deep Learning for Bengali Fake News Detection: Innovative Approaches for Accurate Classification

Sheikh Sadi Bandan., Md Sharuf Hossain., MD. Samiul Islam Sabbir and Khadiza Tul Kobra
Additional contact information
Sheikh Sadi Bandan.: Dept. of Computer Science & Engineering Daffodil International University Dhaka, Bangladesh
Md Sharuf Hossain.: Dept. of Data Science Loyola University Chicago, USA
MD. Samiul Islam Sabbir: Dept. of Computer Science & Engineering Daffodil International University Dhaka, Bangladesh
Khadiza Tul Kobra: Dept. of Information Technology and Management, Illinois Institute of Technology Chicago, USA

International Journal of Research and Innovation in Applied Science, 2024, vol. 9, issue 8, 393-403

Abstract: A vast quantity of data and information are available on the internet. Because the internet is so widely available and has resulted in a tremendous growth in the number of online news, people are interested in reading news from online news portals. Online news portals include things like Facebook, Twitter, WhatsApp, Telegram, Instagram, blogs, and more. Both the quantity of news-on-news websites and the number of readers are increasing. But how real is online news today is a matter of thought. A huge amount of fake news is being spread in newspapers and online due to various yellow journalists. Which is having an adverse effect on society. As a result, there are many kinds of instability, bad politics, etc. problems are being created in the country. If this situation continues, our country and society will go to hell. The only solution is to ensure that yellow journalists do not spread fake news. But despite all the vigilance, fake news will spread. We can solve this by using artificial intelligence, for example, by employing various machine learning and deep learning algorithms, we can identify bogus news and take precautions against it. In this paper, fake news is detected using 4 deep learning algorithms like RNN, LSTM, BiLSTM, GRU model and 1 machine learning algorithm BERT model. RNN has an accuracy of 94.58%, LSTM has an accuracy of 92.84%, BiLSTM has an accuracy of 94.29%, GRU has an accuracy of 93.22% and BERT has an accuracy of 95%. The BERT model has the highest accuracy of 95% among all models.

Date: 2024
References: Add references at CitEc
Citations:

Downloads: (external link)
https://www.rsisinternational.org/journals/ijrias/ ... -issue-8/393-403.pdf (application/pdf)
https://rsisinternational.org/journals/ijrias/arti ... rate-classification/ (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:bjf:journl:v:9:y:2024:i:8:p:393-403

Access Statistics for this article

International Journal of Research and Innovation in Applied Science is currently edited by Dr. Renu Malsaria

More articles in International Journal of Research and Innovation in Applied Science from International Journal of Research and Innovation in Applied Science (IJRIAS)
Bibliographic data for series maintained by Dr. Renu Malsaria ().