Survival Prediction Model of Renal Transplantation using Deep Neural Network

The objectives of this paper are to explore ways to parallelize and distribute deep learning in multi-core and distributed settings. We have heuristically improved the training parameter setting by a Deep Neural Network (DNN) using quad-core CPU and Graphical Processing Unit (GPU) and develop a setting to improve training performances. Along with that, a Parallel Phase Neural network model (PHNNM) has been proposed for the prediction of the long-term survival of liver patients who undergo liver transplantation (LT). We made survival analysis of 13 years in the prediction of liver patients after LT and trained the liver transplantation system to follow up data of 13 years separately using a multilayer perceptron PHNNM model with proper selection of data attributes in conjunction with evaluating the survival probabilities of such data. This paper proved that our prediction model is suitable for the long-term prognosis of survival of patients after LT. The promising results are shown, in combination with the computational performances in terms of CPU and GPU.