HEL-MCNN: Hybrid extreme learning modified convolutional neural network for allocating suitable donors for patients with minimized waiting time

Abstract

Organ transplantation is the ultimate option to treat terminal illness by transplanting the deceased or damaged organs with healthy organs for improving the patient’s life expectancy. The number of organs needed and the organs available for transplantation vary enormously. The tremendous advancements in utilizing big data analytics in the healthcare system make it efficient to explore decision-making information. To make optimal decisions in organ transplantation, this paper proposes a modified convolutional neural network-hybrid extreme learning machine (MCNN-HELM) based prediction model. The proposed MCNN-HELM model utilizes three different real-time datasets as inputs which containrecords of liver, heart, and lung transplantation details of the donor and recipient. At first, the missing values and inaccurate data present in real-time datasets are removed via pre-processing. The pre-processed data are then trained using the MCNN-HELM model that efficiently determines the suitable donor for the recipient by minimizing the waiting time of the recipient for the matching organ donor. Moreover, the MCNN-HELM model gives initial preference to patients with high-risk rates to improve their quality of life. The proposed MCNN-HELM model achieves training accuracy of 97.5% with a computational time of 2.2 s, while the precision value of estimated factual outcomes, potential outcomes, and the accuracy of the best donor type are obtained by 16.3582, 16.1401, and 0.6784 which are more efficient than other state-of-the-art methods.