A Comparative Study of Machine Learning Models for Fake News Detection Using the LIAR Dataset

This This research proposed a machine learning-based framework for fake news detection using the LIAR dataset, incorporating textual and contextual features along with class imbalance handling techniques. Traditional fake news detection approaches primarily focus on textual content and overall accuracy. However, prior studies and experimental observations indicate that such approaches often fail to perform effectively on imbalanced datasets, resulting in biased predictions toward the majority class and poor detection of minority class instances. To address these challenges, the proposed framework introduced: ·                  contextual feature integration (speaker, subject, party, context), TF-IDF with n-gram feature extraction,  class weighting to handle imbalance, comparative evaluation of multiple machine learning models. Experimental evaluation demonstrated that the proposed framework: ·                  improved recall for the minority class (real news), ·                   achieved better balance between precision and recall, ·                  enhanced classification reliability through contextual features, ·                  provided interpretable and efficient model performance. Although Naive Bayes achieved higher accuracy, Logistic Regression with class balancing demonstrated superior overall performance by effectively detecting minority class instances. The results indicate that incorporating contextual information and addressing class imbalance are essential for building reliable fake news detection systems. Therefore, the proposed framework provides a practical and effective approach for real-world fake news classification tasks.  

[1]              William Yang Wang, “Liar, Liar Pants on Fire: A New Benchmark Dataset for Fake News Detection,” Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (ACL), Vancouver, Canada, 2017. [2]              Kai Shu, Amy Sliva, Suhang Wang, Jiliang Tang, and Huan Liu, “Fake News Detection on Social Media: A Data Mining Perspective,” ACM SIGKDD Explorations Newsletter, vol. 19, no. 1, pp. 22–36, 2017. [3]              Hunt Allcott and Matthew Gentzkow, “Social Media and Fake News in the 2016 Election,” Journal of Economic Perspectives, vol. 31, no. 2, pp. 211–236, 2017. [4]              Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova, “BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding,” Proceedings of NAACL-HLT, 2019. [5]              Ashish Vaswani et al., “Attention Is All You Need,” Advances in Neural Information Processing Systems (NeurIPS), 2017. [6]              Fabian Pedregosa et al., “Scikit-learn: Machine Learning in Python,” Journal of Machine Learning Research, vol. 12, pp. 2825–2830, 2011. [7]              David M. Blei, Andrew Y. Ng, and Michael I. Jordan, “Latent Dirichlet Allocation,” Journal of Machine Learning Research, 2003. [8]              Thomas Mikolov et al., “Efficient Estimation of Word Representations in Vector Space,” Proceedings of ICLR Workshop, 2013. [9]              Sepp Hochreiter and Jürgen Schmidhuber, “Long Short-Term Memory,” Neural Computation, vol. 9, no. 8, pp. 1735–1780, 1997. [10]           Christopher D. Manning, Prabhakar Raghavan, and Hinrich Schütze, Introduction to Information Retrieval, Cambridge University Press, 2008.