Multiclass Classification of Liver Diseases using Optimized Machine Learning Classifiers

Chronic hepatitis C is associated with two severe, perhaps fatal health problems: cirrhosis and liver cancer. Chronic liver disease is an exceedingly dangerous condition requiring medical attention. An early liver disease diagnosis is vital for the physician to precisely predict the numerous liver disorders. The proposed research uses the Hepatitis C dataset from the UCI repository for multiclass classification of liver diseases. The pertinent features are carefully chosen using the Minimum Redundancy and Maximum Relevance (MRMR) algorithm. The Synthetic Minority Oversampling Method (SMOTE) was implemented to combat class inequalities. The performance of the model is evaluated using accuracy metrics and the receiver operating curve. The optimizable KNN gave the best performance in terms of training accuracy of 99.41%, test accuracy of 100% with a macro F1 score of 0.982 closely followed by optimizable neural network, optimizable SVM and optimizable ensemble.