Feature Engineering Based Advanced Fake News Detection System Using NLP Techniques

Resumen

The rapid spread of misinformation across digital platforms has made fake news detection an urgent challenge, as deceptive content can significantly influence public opinion and decision-making. Existing approaches often struggle to capture deeper contextual meaning, perform poorly with multilingual content, and lack transparency in explaining their predictions. To address these limitations, this study proposes a novel framework for fake news detection that integrates bidirectional natural language processing, comprehensive feature engineering, and explainable artificial intelligence techniques. For practical application, the system includes a user-friendly interface capable of accepting news content in multiple languages, supported by automatic language detection and translation prior to analysis. The model outputs classification labels along with associated probability scores to improve interpretability. Additionally, it highlights key entities within the text to provide clear explanations for its predictions. Experimental results demonstrate strong performance, achieving a training accuracy of 99.19% and a testing accuracy of 97.338%, indicating the effectiveness and reliability of the approach. Overall, the proposed system offers a highly accurate, multilingual, and interpretable solution for fake news detection.