Fake News Detection Using Machine Learning and NLP Techniques

This paper presented a fake news detection system that uses a five-step NLP preprocessing pipeline combined with TF-IDF feature extraction and four ML classifiers. The preprocessing steps — lowercasing, punctuation removal, stopword removal, tokenisation, and stemming — clean and compress the raw article text before it is turned into a numerical vector. The system architecture was explained using a block diagram and a clear step-by-step description of how data moves through each stage. Among the four classifiers, Logistic Regression performed best, with SVM close behind. Both gave balanced precision, recall, and F1-score values across both the real and fake classes, meaning neither was biased toward one label. Naive Bayes was a strong option when training speed matters. Decision Tree showed clear signs of overfitting and should not be used without ensemble support for this type of task. Ablation testing confirmed that preprocessing genuinely helps; accuracy dropped when steps were removed. The final system is compact, interpretable, and does not require expensive hardware or external data, making it a realistic option for real-world use.

[1]     J. Jouhar and M. Ahmed, “Fake News Detection Using Machine Learning Techniques,” Int. J. Comput. Appl., vol. 15, no. 2, pp. 45–52, 2024. [2]     R. N. Tanaja and P. Kumar, “Machine Learning Models for Fake News Classification,” J. Artif. Intell. Res., vol. 11, no. 1, pp. 101–109, 2025. [3]     O. Embarak and S. Ali, “Deep Learning for Fake News Detection,” Int. J. Adv. Comput., vol. 9, no. 4, pp. 67–75, 2025. [4]     R. K. Ayyasamy and K. Ravi, “Hybrid Deep Learning Framework for Fake News Detection,” IEEE Access, vol. 13, pp. 1100–1110, 2025. [5]     S. Cui and L. Wang, “Fake News Detection Based on Shared Information and Ambiguity Learning,” J. Inf. Process. Syst., vol. 12, no. 3, pp. 211–220, 2026. [6]     K. Shu, A. Sliva, S. Wang, J. Tang, and H. Liu, “Fake News Detection on Social Media: A Data Mining Perspective,” ACM SIGKDD Explor. Newsl., vol. 19, no. 1, pp. 22–36, 2017. [7]     W. Y. Wang, “‘Liar, Liar Pants on Fire’: A New Benchmark Dataset for Fake News Detection,” in Proc. 55th Annu. Meet. ACL, Vancouver, Canada, 2017, pp. 422–426. [8]     C. Castillo, M. Mendoza, and B. Poblete, “Information Credibility on Twitter,” in Proc. 20th Int. Conf. World Wide Web, Hyderabad, India, 2011, pp. 675–684.