Multilingual Fake News Detection

Abstract

Fake news is becoming more of a concern as the number of people who use the internet and have access to the internet continues to expand on a daily basis. The dissemination of misleading information has the potential to cause harm to individuals. In this research work, we offer a system that can automatically rate the possibility that a news article is fraudulent by using machine learning and deep learning techniques. The system is trained on a multilingual dataset consisting of both fake and real news articles from legitimate news websites and datasets from previous works. To make its predictions, the system uses a combination of natural language processing techniques, machine learning, and deep learning algorithms. We examine the performance of the system using a held-out test set and demonstrate its effectiveness in identifying fake news with a high degree of accuracy. The suggested approach has the potential to be an effective instrument in the battle against fake news, contributing to the reduction of the transmission of false information and preventing readers from being misled. In our work SVM, DT, RF, KNN, Logical Regression, NB, XGBoost, and LSTM are applied models. These models are 93%, 88%, 95%, 88%, 94%, 90%, 91%, and 96% accurate. Among all of the algorithms, LSTM has the highest accuracy of 96%.