Hepatitis C Prediction Using Feature Selection by Machine Learning Technique

Abstract

This study suggests a prediction framework for the Hepatitis C virus that is based on machine learning techniques. The authors made use of a dataset available on Kaggle. In this dataset, 564 patients with 12 distinct features are present. They tested two cases, the first one without feature selection and with feature selection based on gain ratio attribute evaluation (GRAE), to guarantee the strength and dependability of the suggested framework. Additionally, an evaluation is conducted on the feature subset that was chosen using the GRAE-generated features. For model evaluation, induction methods and classifiers such as logistic regression (LR), naive bayes (NB), decision tree (DT), support vector machine (SVM), random forest (RF), and multilayer perceptron (MLP) are used. According to the experimental findings, the suggested framework outperformed the others in terms of all accuracy matrices following GRAE selection. According to the experimental findings, the suggested framework outperformed the unfeatured one in terms of accuracy after GRAE selection.