Prediction and Survival Analysis of Patients After Liver Transplantation Using RBF Networks

Prognostic models are becoming useful in assessing the severity of illness and survival analysis in medical domain. Based on the studies, we realized that the current models used in liver transplantation prognosis seems to be less accurate. In this paper, we propose a highly improved model for predicting three month post liver transplantation survival. We performed experiments on the United Nations Organ Sharing dataset, with a 10-fold cross-validation. An accuracy of 86.95% was observed when Radial Basis Function Artificial Neural Network model was used. Other similar methods were compared with the proposed one based on the prediction accuracy. A survival analysis study for a span of 13 years was also done by comparing with MELD an actual dataset. The reported results indicate that the proposed model is suitable for long term survival analysis after liver transplantation.