A COMPARATIVE STUDY OF MACHINE LEARNING ALGORITHMS FOR LEGAL TEXT CLASSIFICATION
DOI:
https://doi.org/10.63503/acset.978-81-995593-9-4.55Keywords:
Legal Text Classification, Machine Learning, TF-IDF, Bayes, Logistic Regression, Support Vector Machine, Natural Language ProcessingAbstract
This paper provides a comparative analysis of machine learning algorithms for text classification in law. The swelling of legal documents and the growing case backlog have led to a demand for automated systems to aid efficient legal analysis. In this article, text data is first preprocessed and converted to numerical features using TF-IDF. Naive Bayes, Support Vector Machine, and Logistic Regression are three machine learning algorithms that are applied and tested on a real-life legal dataset. Evaluation metrics for such models include accuracy, precision, recall, and F1-score. The experimental results indicate that the Support Vector Machine model achieves the highest accuracy among the considered approaches. The results not only confirm that machine learning methods can be applied to large-scale legal text classification but also serve as a reference point for improving the methods in the future through more advanced Naive models.
References
[1] A. Dhar, H. Mukherjee, N. S. Dash, and K. Roy, “Text categorisation: Past and present,” Artificial Intelligence Review, vol. 54, no. 4, pp. 3007–3054, 2021.
[2] L. Wan, G. Papageorgiou, M. Seddon, and M. Bernardoni, “Long-length legal document classification,” arXiv preprint arXiv:1912.06905, 2019.
[3] J. S. T. Howe, L. H. Khang, and I. E. Chai, “Legal area classification: A comparative study of text classifiers on Singapore Supreme Court judgments,” arXiv preprint arXiv:1904.06470, 2019.
[4] K. Kowsari, K. J. Meimandi, M. Heidarysafa, S. Mendu, L. Barnes, and D. Brown, “Text classification algorithms: A survey,” Information, vol. 10, no. 4, p. 150, 2019.
[5] S. Xu, “Bayesian Naïve Bayes classifiers to text classification,” Journal of Information Science, vol. 44, no. 1, pp. 48–59, 2018.
[6] Z. Li, “A classification retrieval approach for English legal texts,” in Proc. 2019 Int. Conf. Intelligent Transportation, Big Data & Smart City (ICITBS), pp. 220–223, 2019.
[7] R. S. Wagh and D. Anand, “A novel approach of augmenting training data for legal text segmentation by leveraging domain knowledge,” in Intelligent Systems, Technologies and Applications, Singapore: Springer, pp. 53–63, 2020.
[8] F. A. Braz, N. C. da Silva, T. E. de Campos, F. B. S. Chaves, M. H. Ferreira, P. H. Inazawa, and F. H. Peixoto, “Document classification using a Bi-LSTM to unclog Brazil's Supreme Court,” arXiv preprint arXiv:1811.11569, 2018.
[9] N. C. da Silva et al., “Document type classification for Brazil’s Supreme Court using a convolutional neural network,” in Proc. 10th Int. Conf. Forensic Computer Science and Cyber Law (ICoFCS), Sao Paulo, Brazil, pp. 29–30, 2018.
[10] A. Iftikhar, S. W. U. Q. Jaffry, and M. K. Malik, “Information mining from criminal judgments of Lahore High Court,” IEEE Access, vol. 7, pp. 59539–59547, 2019.
[11] L. Zhang and D. Moldovan, “Chinese relation classification using long short-term memory networks,” in Proc. 11th Int. Conf. Language Resources and Evaluation (LREC), 2018.
[12] Janani, R., &Vijayarani, S. (2021). Automatic text classification using machine learning and optimisation algorithms. Soft Computing, 25(2), 1129-1145.
