Hindi Fake News Detection Using Machine Learning Models

Authors

  • Varun Rathore, Dr. Latika Jindal

Keywords:

Fake, Machine learning algorithms, Random Forest, CNN, XG Boost

Abstract

Social media and online platforms amplify the spread of misinformation, directly impacting politics, economics, and society. The results for this study will discuss the application of machine learning and deep learning in detecting fake news in Hindi. The final results for this study were derived from a dataset consisting of over 2,100 Hindi news articles labeled as genuine or fabricated. This paper will systematize the results of three robust algorithms: XGBoost, Convolutional Neural Networks (CNN), and Random Forest (RF). These performances were evaluated against performance by doing preprocessing steps such as tokenization, stop word removal, and stemming, with feature extraction using TF-IDF. After all the comparisons, the best performance was given by XGBoost with 94% accuracy, beating Random Forest with an accuracy of 85% and CNN with an accuracy of 84%. Moreover, XGBoost beat on other metrics too, such as RMSE and MAE. The above findings have underlined the strong potential of both ensemble and deep learning models in the detection of fake news in Hindi.

Downloads

Download data is not yet available.

Metrics

Metrics Loading ...

Downloads

Published

2025-04-25

How to Cite

1.
Varun Rathore, Dr. Latika Jindal. Hindi Fake News Detection Using Machine Learning Models. J Neonatal Surg [Internet]. 2025Apr.25 [cited 2025Sep.22];14(18S):299-305. Available from: https://www.jneonatalsurg.com/index.php/jns/article/view/4646