Fake News Detection Using Machine Learning

The rapid spread of fake news on social media and online platforms has become a major challenge to information credibility. This paper proposes a machine learning-based approach for detecting fake news using textual features. A real-world dataset containing 44,942 news articles (21,418 true and 23,524 fake news) is utilized for training and testing. Multiple classifiers, including Logistic Regression, Decision Tree, Random Forest, and Gradient Boosting, are implemented to enhance prediction reliability. The system preprocesses the data, converts text into numerical feature vectors, and evaluates model performance using standard metrics such as accuracy, precision, recall, and F1-score. Experimental results indicate that the proposed system effectively distinguishes between true and fake news, providing a scalable solution for real-world applications. A comparative analysis of different machine learning models is conducted to identify the most effective classifier for fake news detection. The results demonstrate that ensemble-based models achieve superior performance compared to individual classifiers. The proposed approach can be effectively integrated into online news platforms and social media systems to reduce the spread of misinformation and improve information authenticity