Efficient vs Transformer-Based Misinformation Detection: A Comparative Study of PAC and RoBERTa
DOI:
https://doi.org/10.32792/universityofthi-qar.v21i2.495Keywords:
Misinformation Detection,Passive-Aggressive Classifier, Fake News Classification, RoBERTa, Transformer-based, Error AnalysisAbstract
Misinformation is one of the most important factors to consider in this day and where information dissemination is common, and misinformation can impact people’s emotions and decisions and affect societal peace. Even though transformers have proven to be quite useful, they require large amounts of processing power, posing a problem in terms of scalability. The present study solves this issue by performing an objective analysis, which is aimed at comparing the passive-aggressive classifier algorithm as the classical machine learning method with the RoBERTa transformer-based model with an equal experimental environment. For conducting the experiment, the researchers use the labeled data set consisting of 44,898 news pieces with fake and real samples. A uniform preprocessing procedure and stratified data splitting were used to ensure fairness and replicability. While PAC learned on TF-IDF features with n-grams representation, RoBERTa was fine-tuned on contextual embeddings inside a transformer-based framework. The research showed that RoBERTa provides slightly better prediction accuracy. Still, this difference is not statistically significant according to the McNemar test. At the same time, the PAC proved to be far more computationally efficient, requiring considerably less time and space resources for training. As a result, RoBERTa was more suitable for applications involving real-time data processing and limited resources availability. Thus, small-scale machine learning algorithms can reach nearly the same level of performance as transformer-based models if optimized correctly.
References
A. Sarkar, A. B. Chowdhury, and M. N. B, “Classification of Online Fake News Using N-Gram Approach and Machine Learning,” vol. 2, pp. 322–336, 2023.
E. Comito, C., Caroprese, L. & Zumpano, “Multimodal fake news detection on social media: a survey of deep learning techniques,” Soc. Netw. Anal. Min., vol. 13, 2023.
J. Alghamdi, S. Luo, and Y. Lin, “A comprehensive survey on machine learning approaches for fake news detection,” pp. 51009–51067, 2024, doi: 10.1007/s11042-023-17470-8.
M. F. Lazuardi and R. Hiunarto, “Hoax News Detection Using Passive Aggressive Classifier and TfidfVectorizer,” J. Tek. Inform., vol. 16, pp. 185–193, 2023.
O. D. Okey, E. U. Udo, R. L. Rosa, D. Z. Rodríguez, and J. H. Kleinschmidt, “Investigating ChatGPT and cybersecurity: A perspective on topic modeling and sentiment analysis,” Comput. Secur., vol. 135, p. 103476, 2023, doi: https://doi.org/10.1016/j.cose.2023.103476.
A. Saeed and E. Al Solami, “Fake News Detection Using Machine Learning and Deep Learning Methods,” 2023, doi: 10.32604/cmc.2023.030551.
G. Chhetri, “WISE?: Web Information Satire and Fakeness Evaluation,” pp. 93–102.
G. Airlangga, “Comparative Analysis of Machine Learning Algorithms for Detecting Fake News: Efficacy and Accuracy in the Modern Information Ecosystem,” J. Comput. Networks, Archit. High Perform. Comput., vol. 6, pp. 354–363, 2024.
S. W. S. and W. W. S. F. N. Azizah, H. D. Cahyono, “Performance Analysis of Transformer Based Models (BERT, ALBERT, and RoBERTa) in Fake News Detection,” 6th Int. Conf. Inf. Commun. Technol. (ICOIACT), Yogyakarta, Indones., pp. 425–430, 2023.
S. P. Kasiviswanathan and S. S. S. Kumar, “Impact of TF-IDF and N-Gram Features on Passive-Aggressive Classifier,” IJACSA, vol. 14, no. 5, 2023.
R. Khan et al, “Detection of fake news using Naive Bayes and Passive Aggressive Classifier,” IJSREM, vol. 9, no. 14, 2025.
R. Khan, S. K. Mishra, R. Kumari, and S. Sinha, “Research paper on Detection of fake news using Naive Bayes and Passive Agressive Classifier,” pp. 1–8, 2023, doi: 10.55041/IJSREM22437.
N. Rishitha and D. K. Reddy, “A Web-Based Application for Fake News Detection,” no. October, 2025.
M. Nadeem, “Enhancing Fake News Detection with a Hybrid NLP-Machine Learning Framework. International Journal of Advanced Research in Computer and Communication Engineering,” Int. J. Adv. Res. Comput. Commun. Eng., vol. 13, pp. 203–214, 2024.
S. Kumari, “A Deep Learning Multimodal Framework for Fake News Detection,” vol. 14, no. 5, pp. 16527–16533, 2024.
T. Jiang, J. P. Li, A. U. Haq, A. Saboor, and A. Ali, “A Novel Stacking Approach for Accurate Detection of Fake News,” IEEE Access, vol. 9, pp. 22626–22639, 2021, doi: 10.1109/ACCESS.2021.3056079.
B. M. Merzah, J. Razmara, and Z. Salmanian, “Hybrid deep learning models for fake news detection?: case study on Arabic and English languages,” no. January, pp. 1–19, 2026, doi: 10.3389/fdata.2025.1683786.
M. Nadeem, “Enhancing Fake News Detection with a Hybrid NLP-Machine Learning Framework,” vol. 1, no. 3, pp. 203–214, 2024.
et al. Y. Liu, “RoBERTa: A Robustly Optimized BERT Pretraining Approach,” Prepr. arXiv, vol. 9, no. 1, 2019.
et al. V. S. Chauhan, “Sentiment Analysis and Fake News Detection using Machine Learning Algorithms (PAC, NB, SVM),” ResearchGate, vol. 0, no. 1, 2024.
et al. H. Ahmed, “Detection of Fake News using N-Gram Analysis and PAC Classifier,” Comput. Human Behav., vol. 145, 2023.
V. Perez-Rosas et al., “Real-time PAC Applications for Twitter Misinformation,” Digit. Investig. J., vol. 42, no. 11, 2025.
S. Tyagi and P. Kumar, “Hybrid Probabilistic and Linear Models for Robust News Classification,” Expert Syst. Appl., vol. 2, no. 2, 2024.
K. S. Jones, “Practical Frameworks for Deploying Lightweight Fake News Systems,” Softw. Pract. Exp., vol. 22, no. 1, 2025.
M. Gupta et al., “Linguistic Feature Comparison across Classical ML Models,” J. Big Data, vol. 11, 2023.
R. J. Mooney et al., “Hybrid ML Techniques for Scalable Text Classification,” Inf. Process. Manag., vol. 7, no. 1, 2025.
L. Wu and H. Liu, “A Comprehensive Comparison of ML and DL paradigms for Misinformation,” ACM Comput., vol. 56, no. 4, 2024.
et al R. K. Kaliyar, “DeepFake: Improving Fake News Detection using Deep Learning Models (CNN and GPT),” Soft Comput., vol. 120, 2023.
K. Sharma and R. Kapoor, “Performance Analysis of BERT, RoBERTa, and XLNet in Fake News Detection,” J. Inf. Sci., vol. 3, 2024.
J. Devlin and M. Chang, “Enhancing Sequence Modeling for Fake Content Detection using RoBERTa-LSTM,” J. Mach. Learn. Res., vol. 22, no. 2, 2024.
Z. Akata et al., “Multimodal Learning for Richer Representation in Fake News,” IEEE Signal Process. Mag., vol. 22, no. 4, 2026.
J. Zhou et al, “Graph Neural Networks for Semantic Relationship Modeling in Misinformation,” IEEE AI Rev., vol. 3, no. 4, 2022.
T. Chen and C. Guestrin, “Hybrid News Classification: Integrating Transformers with XGBoost,” J. Comput. ACM SIGKDD, vol. 22, no. 1, 2024.
Published
How to Cite
Issue
Section
