Machine Learning Models for Fake News Detection: A Comparative Analysis

Abstract

In the context of this paper, the focus has been on considering the use of machine learning models for fake news which is a growing concern in the growing digital age. The study focuses on four models of the model Random Forest, AdaBoost, Support Vector Machine, and Long Short-Term Memory to assess the accuracy of classification of the news articles as fake or actual. A fairly considerable volume of data was chosen to train and validate the models, and their efficiency was evaluated with metrics of classification accuracy. The findings show that the model we used for classification, Random Forest, achieved the highest level of accuracy, 76.43% when used to differentiate between fake and real articles from the other three models. Thus, the findings indicate that ML, especially Random Forest, can serve as a useful instrument used to counterfactual the phenomenon of fake news and misinformation by developing an automatic classification method for news articles. The following has been pointed out that these models can benefit media, social networks, and independent checkers as different sides of the information war against the issue. Therefore, the research advances the existing knowledge in the area of the development of automated systems for the classification of fake news.