A data-driven framework for fair and efficient organ transplantation using gradient boosting and adaptive genetic allocation

Abstract

This study proposes a comprehensive data-driven framework aimed at enhancing organ transplantation efficiency through optimized risk assessment, donor-recipient matching, and equitable organ allocation. Utilizing the gradient boosting algorithm (GBA) for risk prioritization, A* search for optimal donor location, the modified convolutional neural network-based hybrid extreme learning classifier (MCNN-HELM) model for precise matching, and an adaptive objective-weighted genetic allocation (AOWGA) algorithm, the framework addresses critical challenges in organ allocation and distribution. The experimental results indicate strong performance metrics, with the integrated system achieving an overall accuracy of 96%, allocation efficiency of 97%, and a fairness index of 0.92. The MCNN-HELM model showed a matching precision of 0.94 and an accuracy of 97.5%, outperforming existing methods. AOWGA surpassed comparative allocation methods, demonstrating an allocation efficiency of 0.96 and a positive outcome rate of 0.95. By integrating these modules, the framework not only improves organ allocation processes but also enhances survival rates and promotes ethical practices in organ distribution. By innovatively integrating these techniques, the framework reduces waiting times, improves patient outcomes, and ensures fair allocation, marking a significant advancement in addressing the persistent organ shortage and setting a new standard for ethical and efficient organ transplantation.

Keywords: Donor–recipient matching and allocation; Organ transplantation; Prioritization; Risk assessment.