Prediction of Liver Diseases Using Neural Network Analysis

Liver is a vital organ in the body and works to filter blood from the digestive tract before passing it on to the rest of the body. Liver diseases are varied and may be assessed by liver function tests including ALT. The main objectives of this study were to use neural network analysis to predict liver disease, and to identify the relative contribution of liver disease predictors. A dataset of Indian liver patients posted on Kaggle was used to be analyzed for liver disease prediction. The dataset included 583 subjects among whom 71.4% had liver disease. Study predictors included age, gender, ALT, AST, bilirubin, albumin, total protein, albumin/globulin ratio, and alkaline phosphatase. The prediction model was effective in 79.6% predicting the liver disease. The most important predictor was ALT, and the least important predictor was alkaline phosphatase. Taken together, using neural network analysis is effective in predicting liver disease from one side and from another side, it can be improved to give more accurate results.