Citation

Das, Nipa Rani and Md., Mahbubul Alam and Polok, Asif Pervez and Raquib, Mirza and Al Mamun, Abdullah Sarwar and Hossen, Md. Jakir (2024) Social Media Bangla Fake News Detection Using Deep and Machine Learning Algorithms. International Journal of Engineering Trends and Technology, 72 (5). pp. 346-354. ISSN 2231-5381

Text Social Media Bangla Fake News Detection Using Deep and Machine Learning Algorithms.pdf - Published Version Restricted to Repository staff only Download (825kB)

Abstract

In this current world, media and online news publications are spreading rapidly. The dissemination of inaccurate information on social media platforms is increasingly becoming a concern. The proliferation of fake news is due to the ease with which data can be accessed and shared as a result of the mobile technology revolution. Like many countries, fake news is spreading very fast in Bangladesh. The situation gets worse with the spreading of misinformation about epidemics like COVID19. Created a novel dataset of the Bengali language and achieved to goal by using LSTM and machine learning models. Now, other algorithms are used, but the LSTM and machine learning models have good performance. This program's algorithm to select the attribute, a text feature based on TF-IDF and Word Embedding was used. Focused LSTM-base model and machine learning models, especially the Bangla-LSTM-base model and machine learning models. Finally, add a dense layer as a summary layer responsible for generating summary sentences to text. According to all of the evaluations performed above, the additional Trees Classifier outperformed the other six Machine Learning methods. The accuracy rate for identifying false news in news headline data is roughly 86.14%. The second-best accuracy provided by the Random Forest Classifier algorithm is close to 85%. The third-best accuracy provided by the Decision Tree Classifier method is approximately 84%. Moreover, seeing that deep learning algorithms outperformed machine learning ones. Furthermore, LSTM has a 96.14% training accuracy rate and an 86% testing accuracy rate for identifying false news in news headline data

Item Type:	Article
Uncontrolled Keywords:	LSTM, Word Embedding, Dataset, Headline, TF-IDF, Accuracy
Subjects:	Q Science > QA Mathematics > QA71-90 Instruments and machines > QA75.5-76.95 Electronic computers. Computer science T Technology > T Technology (General)
Divisions:	Faculty of Engineering and Technology (FET)
Depositing User:	Ms Nurul Iqtiani Ahmad
Date Deposited:	03 Jul 2024 02:51
Last Modified:	03 Jul 2024 02:51
URII:	http://shdl.mmu.edu.my/id/eprint/12583

Downloads

Edit (login required)