Volume 18, No. 6, 2021
Design Machine Learning Algorithms To Predict Liver Transplantation Survival
Prof. Bere Sachin Sukhadeo, Prof. Salunke Shrikant Dadasaheb, Prof. Bhange Rajani Rahul , Prof. Zol Ramdas Madhukar , Prof. Shende Sachin Santosh
Abstract
Liver transplantation (LT) is an important therapy option for people with liver disease, which has recently been improved by computerized medical field technology. Patients with LT have a bad prognosis in some cases, which is the main concern in many settings. A variety of predictive models have been developed by academics to address these difficulties. As a result, the current focus of research is on developing more precise and accurate prediction methods for use with advanced MLP techniques. Predictions are made in multiple stages using the method described below. The United Nations Organ Sharing database is used to gather medical data in the early stages of the transplant process (UNOS). Specifically, we drew on the UNOS for information on liver disease. For simplicity, the data are fed into the principal component analysis (PCA) to reduce the dimensions of the attributes. Using the Advance MLP categorization, patients will have a better chance of surviving LT if they fall into one of two distinct categories: "best survival" or "worst survival." The suggested framework was subjected to a stimulation study in order to determine its performance. There are many variables that are estimated for this proposed framework such as the following: precision/accuracy/specificity/error/f1 score/fpr/kappa/MCC. The proposed framework has an accuracy rating of 98%, which proves that the proposed design is accurate.
Pages: 9598-9607
Keywords: Artificial Neural Networks, Liver Transplantation, Machine Learning, Multilayer Perceptron, Post-Liver Transplantation Survival Prediction.