Citation :

Enas Mohamed Ali Quteishat, Ahmed Qtaishat, "Voting-Based Ensemble Learning Framework for Explainable Supreme Court Decision Prediction," International Journal of Electronics and Communication Engineering, vol. 13, no. 6, pp. 98-111, 2026. Crossref, https://doi.org/10.14445/23488549/IJECE-V13I6P108

Abstract

Prediction of judicial decisions is a growing application of Artificial Intelligence (AI) due to the increasing volume of legal documents and judicial case records being maintained, combined with the increasing complexity associated with developing consistency in data-driven decision-making. Previous studies used machine learning algorithms and Deep Learning (DL) techniques to predict the outcomes of courts, and most of the existing techniques rely on single modeling approaches that are sensitive to data imbalance, noise, and the ever-changing nature of judicial activity across time. As a result, this study will address these limits by creating an ensemble model based on an ensemble voting method for predicting the case outcomes of the Supreme Court of the United States. The dataset used for this study will be over 3,304 unique Supreme Court decisions made from 1955 through to 2021. Legal cases contained in the dataset will be converted into numerical representations of each legal document using the TF-IDF method, and then a combination of seven different heterogeneous classifiers will be developed and combined in an ensemble model through a soft-voting strategy. The ensemble model developed proved to have a significantly higher accuracy score than any of the baseline models individually, achieving a level of accuracy of 92.5% (ROC: 0.91), which demonstrates the strong predictive capability of the ensemble model and its ability to differentiate between the different outcome categories. To improve transparency and reduce potential ethical concerns with AI judicial systems, an SHAP-based method was utilized to generate interpretability and explainability with respect to the analysis of the factors that contribute to each model's predictions. The interpretability of the model identified three major categories: the issue area, the type of litigant, and how the lower Court decided the Case, as key features that appear to influence the predictions generated by the model. The additional tests confirmed that, regardless of the judicial time, ideology, and split time, the overall performance of the ensemble model was stable and showed only slight decreases in performance. Ultimately, this research indicates that robust, transparent, and ethically based methods to predict outcomes of legal rulings will be possible by merging Ensemble Learning with XAI (Explanatory Artificial Intelligence) techniques to create and implement Legal Analytics as a systematic process for assisting Legal Judgements.

Keywords

Artificial Intelligence, Machine Learning, Deep Learning, Natural Language Processing, Judicial Decision Prediction.

References

1. Andrew D. Martin et al., “Competing Approaches to Predicting Supreme Court Decision Making,” Perspectives on Politics, vol. 2, no. 4, pp. 761-767, 2004.
   [CrossRef] [Google Scholar] [Publisher Link]
2. Daniel Martin Katz, Michael J. Bommarito, and Josh Blackman, “A General Approach for Predicting the Behavior of the Supreme Court of the United States,” PloS one, vol. 12, no. 4, pp. 1-18, 2017.
   [CrossRef] [Google Scholar] [Publisher Link]
3. Pauline Kim et al., “Supreme Court Forecasting Project: Legal and Political Science Approaches to Supreme Court Decision-Making,” Columbia Law Review, vol. 104, pp. 1150-1210, 2004.
   [Google Scholar] [Publisher Link]
4. Daniel Martin Katz, Michael J Bommarito, and Josh Blackman, “Predicting the Behavior of the Supreme Court of the United States: A General Approach,” arXiv preprint, pp. 1-17, 2014.
   [CrossRef] [Google Scholar] [Publisher Link]
5. Supreme Court Judgment Prediction. [Online]. Available: https://www.kaggle.com/datasets/deepcontractor/supreme-court-judgment-prediction
6. Aaron Russell Kaufman, Peter Kraft, and Maya Sen, “Improving Supreme Court Forecasting Using Boosted Decision Trees,” Political Analysis, vol. 27, no. 3, pp. 381-387, 2019.
   [CrossRef] [Google Scholar] [Publisher Link]
7. Sugam K. Sharma, Ritu Shandilya, and Swadesh Sharma, “Predicting Indian Supreme Court Judgments, Decisions, or Appeals: eLegalls Court Decision Predictor (eLegPredict),” Statute Law Review, vol. 44, no. 1, pp. 1-9, 2023.
   [CrossRef] [Google Scholar] [Publisher Link]
8. Mohammad Alali et al., “JUSTICE: A Benchmark Dataset for Supreme Court's Judgment Prediction,” arXiv preprint, pp. 1-6, 2021.
   [CrossRef] [Google Scholar] [Publisher Link]
9. Joseph A. Ignagni, “Explaining and Predicting Supreme Court Decision Making: The Burger Court's Establishment Clause Decisions,” Journal of Church and State, vol. 36, no. 2, pp.  301-327, 1994.
   [CrossRef] [Google Scholar] [Publisher Link]
10. N. Sivaranjani, and J. Jayabharathy, “Forecasting the Decision Making Process of Supreme Court using Hierarchical Convolutional Neural Network,” International Journal of Ad Hoc and Ubiquitous Computing, vol. 40, no. 1-3, pp. 116-126, 2022.
    [CrossRef] [Google Scholar] [Publisher Link]
11. Sugam Sharma, Ritu Shandilya, Swadesh Sharma, “Predicting Indian Supreme Court Judgments, Decisions, or Appeals,” ArXiv Preprint, pp. 1-9, 2021.
    [CrossRef] [Google Scholar] [Publisher Link]
12. Kimo Gandall et al., “Predicting Precedent: A Psycholinguistic Artificial Intelligence in the Supreme Court,” Case Western Reserve: Journal of the Law, Technology & the Internet, vol. 14, no. 2, pp. 1-55, 2023.
    [Google Scholar] [Publisher Link]
13. Daniel Martin Katz, Michael James Bommarito, and Josh Blackman, “Crowdsourcing Accurately and Robustly Predicts Supreme Court Decisions,” arXiv preprint, pp. 1-11, 2017.
    [CrossRef] [Google Scholar] [Publisher Link]
14. M. Salman Abbasi et al., “Leveraging Autocorrelation in a Dilated CNN-LSTM Framework for Predicting the US Supreme Court Decisions,” IEEE Access, vol. 13, pp. 161250-161261, 2025.
    [CrossRef] [Google Scholar] [Publisher Link]
15. Kankawin Kowsrihawat, Peerapon Vateekul, and Prachya Boonkwan, “Predicting Judicial Decisions of Criminal Cases from Thai Supreme Court Using Bi-directional GRU with Attention Mechanism,” 2018 5^th Asian Conference on Defense Technology (ACDT), Hanoi, Vietnam, pp. 50-55, 2018.
    [CrossRef] [Google Scholar] [Publisher Link]
16. Susan S. Silbey, “The Dream of a Social Science: Supreme Court Forecasting, Legal Culture, and the Public Sphere,” Perspectives on Politics, vol. 2, no. 4, pp. 785-793, 2004.
    [Google Scholar] [Publisher Link]
17. Leighton Vaughan Williams, “Forecasting the Decisions of the US Supreme Court: Lessons from the ‘Affordable Care Act’Judgment,” The Journal of Prediction Markets, vol. 9, no. 2, pp. 64-78, 2015.
    [CrossRef] [Google Scholar] [Publisher Link]
18. Ali S. Masood, and Donald R. Songer, “Reevaluating the Implications of Decision-Making Models,” Journal of Law and Courts, vol. 1, no. 2, pp. 363-389, 2013.
    [CrossRef] [Google Scholar] [Publisher Link]
19. Emre Mumcuoğlu et al., “Natural Language Processing in Law: Prediction of Outcomes in the Higher Courts of Turkey,” Information Processing & Management, vol. 58, no. 5, pp. 1-16, 2021.
    [CrossRef] [Google Scholar] [Publisher Link]
20. Michael Benedict L. Virtucio et al., “Predicting Decisions of the Philippine Supreme Court Using Natural Language Processing and Machine Learning,” 2018 IEEE 42^nd Annual Computer Software and Applications Conference (COMPSAC), Tokyo, Japan pp. 130-135, 2018.
    [CrossRef] [Google Scholar] [Publisher Link]
21. Octavia-Maria Sulea et al., “Predicting the Law Area and Decisions of French Supreme Court Cases,” arXiv preprint, pp. 1-7, 2017.
    [CrossRef] [Google Scholar] [Publisher Link]
22. Nikolaos Aletras et al., “Predicting Judicial Decisions of the European Court of Human Rights: A Natural Language Processing Perspective,” PeerJ Computer Science, vol. 2, pp. 1-19, 2016.
    [CrossRef] [Google Scholar] [Publisher Link]