Artificial Neural Networks Model in the Prediction of Pancreas Transplantation Recipient Survival

Abstract

The pancreas transplantation is always a risky procedure, so that the number of transplantation suggesting by doctors are less. Also doctors have no method to predict the survival of patient after pancreas transplantation. So it is necessary to have a system to predict the survival of patients after pancreas transplantation. Mostly Type 1 diabetes are treated through pancreas transplantation. Yet, it is commonly held for those with genuine diabetes difficulties, on the grounds that the reactions of a pancreas transplant are huge. Consequently the requirement for this investigation is basic. Pancreas transplantation is constantly dangerous in light of it’s eventual outcomes, so the proposed framework anticipating the odds of survival of patients by breaking down their wellbeing conditions. This investigation will be useful to specialists for understanding careful results and expands the believability of specialists. From our research, we can see that the chosen model have given greater survival rate. Pancreas transplantation is definitely not a standard procedure yet a seldom happening procedure, since it is an exceptionally unsafe technique and has numerous eventual outcomes. Need an exact model to foresee the survival after transplantation. We proposed three models which are suitable for pancreas survival prediction. From that, with Multilayer Perceptron we obtained 99.5% accuracy in survival prediction. The result was compared with existing research work to ensure that it is the only accurate method now existing to predict the patients survival in pancreas transplantation.