The Fight Against Fake News Leveraging Machine Learning, NLP as Deterring Methods

Fake news spreads rapidly and discreetly, influencing public opinion and decision-making with potentially significant consequences. Detecting misinformation before it proliferates is therefore crucial to mitigate its impact, yet manual fact-checking is impractical and systems relying on constantly updated information face challenges in scalability. Therefore, this research paper proposes an automated approach to flag misleading articles based on linguistic patterns rather than factual accuracy. For this purpose, five supervised machine learning (ML) algorithms were tested using natural language processing (NLP) on the ISOT Fake News Dataset, comprising 44,897 articles. Among them, Logistic Regression (LR) achieved the highest accuracy of 96.34%, demonstrating that even simpler models can effectively distinguish between fake and genuine news. These findings highlight the potential of NLP-based ML models in combating misinformation in real time.