A Hybrid Diagnosis Model to Predict Liver Disease Using Machine Learning Techniques

Abstract

The medicare business produces a vast number of records. We realize that techniques for machine learning (ML) may also be utilised to locate covert data for diagnostic purposes and constructive decision making. Liver diseases have quickly grown in modern days and in many other nations like China, Moldava etc. are regarded a highly deadly condition. The major goal is to forecast liver diseases utilising various categorization methods for this study article. The logistical regression (LR), Random Forest (RF) K-Nearest neighbour (KNN), Naïve Bayes (NB), Decision Tree (DT) and support vector machine (SVM) are the methods utilised for this research. To evaluate this classification method, efficiency rating and confusion matrix are utilised. Liver disease is a general clinical problem linked to many discomforts and severe death rate. It is of fundamental importance that disease be detected before so much may be saved. The hepatitis stages represent an important point of view for targeted therapy. The therapeutic experts are incredibly difficult about predicting the sickness in its early days because of sensitivity events. In particular, after the point of no return, the negative impacts become obvious. We have liver issue forecasts to overcome this problem. Liver disease may be identified by unimaginable control systems, and the usage prediction for a number of features and classification blending has been categorised. We used 6 kinds of NB, LR, SVM, RF, KNN and DT to examine liver disease in this study. The exhibits are evaluated with 5 different performance metrics: accuracy, kappa, absolute mean error (MAE), root average square error (RMSE) and F-score. The aim of this inquiry study is to predict and identify the most effective method of liver disease using ML.