Classification and Prediction of Liver Disease Diagnosis Using Machine Learning Algorithms

Abstract

Liver diseases which is a chronic disease that lasts more than six months are one of the most dangerous and sounding alarms in the health care systems of the world due to the prediction of its enhancement due to several factors such as an increase in the consumption of alcohols, deteriorating polluting the situation in the whole world due to global warming and heavy industrialization and exhaust of toxic gases, contaminated water, and food, drug, primarily poor lifestyle choices lead to continuous increase in the diagnosis of anomalies in the liver of the Patients. The patient’s liver datasets are explored to build classification and prediction models for early diagnosis of liver disease. In an effort to reduce the workload on doctors, machine learning is used to predict disease. This paper explores many historical machine-learning models for liver diseases diagnosis and classification. The comparative evaluation of more than six models suggests the best method for the targeted dataset. Various ensemble techniques and tunning of hyperparameters suggest that these techniques may result in better accuracy but with the increased cost of computing, efficiency makes them irrelevant for real-world applications for offline problems these are the best bet for enhanced accuracy of the model.