Liver Cirrhosis Stage Prediction Using Machine Learning: Multiclass Classification

The chapter delves into the application of machine learning algorithms for predicting liver cirrhosis stages, a critical health issue affecting millions worldwide. It highlights the use of seven machine learning algorithms—Random Forest, Logistic Regression, Support Vector Machine, K-Nearest Neighbors, Decision Tree, Naive Bayes, and Artificial Neural Network—to classify patients into four stages of liver cirrhosis. The study focuses on multiclass classification on an imbalanced dataset, employing feature selection techniques such as Random Forest and a combination of Random Forest with Mutual Information. The chapter compares the performance of these algorithms using metrics like F1 score and AUC-ROC, showcasing the superior performance of the Artificial Neural Network in terms of AUC-ROC. The research also proposes a novel feature selection method combining Random Forest and Mutual Information, demonstrating its effectiveness in improving classification performance. The chapter concludes with recommendations for future research, including the use of larger datasets and further validation of the proposed feature selection method.