Authors:
P. Kiruthiga, S. Silvia Priscila
Addresses:
Department of Computer Applications, Bharath Institute of Higher Education and Research, Chennai, Tamil Nadu, India. Department of Computer Science, Bharath Institute of Higher Education and Research, Chennai, Tamil Nadu, India.
Abstract:
The proposed study presents a new Explainable Artificial Intelligence solution that optimises the match between kidney donors and recipients and forecasts post-transplant survival rates. Conventional matching algorithms are typically opaque systems, where medical practitioners have no idea of the logic behind a particular match. To address this, researchers apply an advanced ensemble learning methodology, combined with local interpretable model-agnostic explanations, to enhance transparency in clinical decision-making. The dataset used in this study is a specialised one comprising 432 clinical cases and various physiological and immunological parameters. The Python-based environments were used for data processing and model development, using libraries such as Scikit-Learn for prediction and SHAP for explanation. The framework, by placing greater emphasis on interpretability, has identified the determinants of survival, including age compatibility and human leukocyte antigen matching. The findings demonstrate that the proposed system not only ensures high predictive value for graft survival but also provides practical information that increases clinicians' confidence. Such a twofold interest in performance and transparency is one of the main advances in medical informatics, and both high-quality data and transparent, logical arguments support life-critical transplant decisions.
Keywords: Explainable AI; Kidney Transplantation; Survival Prediction; Donor Matching; Clinical Decision Support; Conventional Matching; Clinical Decision-Making; Performance and Transparency.
Received on: 21/04/2025, Revised on: 14/08/2025, Accepted on: 11/11/2025, Published on: 07/06/2026
DOI: 10.64091/ATIHL.2026.000305
AVE Trends in Intelligent Health Letters, 2026 Vol. 3 No. 2 , Pages: 116-126