Hepatitis Disease Prognosis Using Random Forest, K-Nearest Neighbors, Naive Bayes, Support Vector Machine, and Multi-Layer Perceptron

Abstract

Among the leading infectious killers globally right now is hepatitis. In humans, it causes inflammation of the liver.
We have a great opportunity to save many lives by detecting this fatal condition early on. In this study, we used
several data mining approaches to forecast the occurrence of hepatitis. In addition to this, we have put out a
respectable method for enhancing the accuracy of our prediction models. We eliminated observations with missing
values as a means of dealing with missing data in our dataset. Using the info-gain feature selection technique in
conjunction with ranker search, we were able to identify the characteristics that were not needed. The hepatitis
illness dataset is used to determine the prediction accuracy using classification approaches such as K-Nearest
Neighbors (KNN), Naive Bayes Support Vector Machine (SVM),