Authors : Vidyasri G.; Dr. K. Priya

Volume/Issue : Volume 11 - 2026, Issue 10 - October

DOI : https://doi.org/10.38124/ijisrt/26Oct569

Note : A published paper may take 4-5 working days from the publication date to appear in PlumX Metrics, Semantic Scholar, and ResearchGate.

Abstract : Online news services and social networks enable information to circulate rapidly, but they also make it easier for inaccurate stories to reach large audiences. Checking every item manually is difficult when the volume of published content is high. This paper examines a text-classification pipeline that labels news articles as FAKE or REAL. Article titles and body text are cleaned using Natural Language Processing (NLP), represented as Term Frequency–Inverse Document Frequency (TF-IDF) vectors, and passed to a Logistic Regression classifier. Performance is assessed with accuracy, precision, recall, F1-score, and a confusion matrix. On the selected test set, the model obtained 75.85% accuracy and a 77.95% F1-score. The results illustrate both the usefulness and the limitations of a lightweight, text-based approach: the classifier identified REAL articles more reliably than FAKE articles.

Keywords : Fake News Classification, NLP, TF-IDF, Logistic Regression, Text Mining, Machine Learning.

References :

1. J. Alghamdi, S. Luo, and Y. Lin, “A comprehensive survey on machine learning approaches for fake news detection,” Multimedia Tools and Applications, vol. 83, pp. 51009–51067, 2024.
2. “Fake News Detection Using Natural Language Processing and Machine Learning Techniques,” 2023 Seventh International Conference on Image Information Processing (ICIIP), IEEE, 2023, doi: 10.1109/ICIIP61524.2023.10537640.
3. R. Hudaini, M. Hendriansyah, and E. Gautama, “Fake News Classification Using Logistic Regression with TF-IDF Feature Extraction,” Perbanas International Seminar on Economics, Business, Management, Accounting and IT, 2026.