Investigating the Efficacy of Artificial Intelligence and Natural Language Processing for Fake News Detection

Abstract

The fast growth of internet platforms and social media has led to in an incredible flood of information, making it increasingly difficult to distinguish between authentic news and fake content. Fake news poses serious threats to public debate, and societal stability. To solve this issue, researchers used machine learning and natural language processing (NLP) approaches to create automated systems for detecting fake news. This paper provides an overview of cutting-edge strategies for detecting false news using machine learning and NLP. We investigate text preparation techniques such as tokenization, stemming, and stop-word removal to prepare data for analysis. We investigated our findings using a variety of machine learning algorithms, including logistic regression, decision trees, support vector machines, random forests, gradient boosting, and neural networks. We demonstrate that the combination of machine learning and NLP techniques provides fascinating possibilities for countering the spread of fake news. We were able to improve our ability to detect deceptive content. Random Forest (RF), CatBoost, and XG Boost are among the best-performing algorithms. When applied to the Fake News Net dataset, the Random Forest method demonstrated exceptional performance inside experiments using a partitioning ratio of 70-30 for training and testing, respectively, with a stunning accuracy of 99.02%.