Currently in the world of medicine, determining liver inflammation is something that is not easy to do. But there are medical records that have kept the patient's symptoms and diagnosis of liver inflammation. The weaknesses of the manual method encourage researchers to develop a method that does not depend 100% on humans. The developed method utilizes a computer as a tool to analyze data. This kind of thing is certainly very useful for health experts. They can use existing medical records as an aid in making decisions about the diagnosis of a patient's disease. In this study, we analyzed the performance of machine learning algorithms by comparing the support vector machine, naïve Bayes and k-nearest neighbor algorithms. This study aims to determine the performance of which algorithm has the highest accuracy in liver disease data. From the research results using splinting data 80:20 it can be concluded that the Naïve Bayes algorithm model has better performance than other algorithm models when using the SMOTE technique with an accuracy value of 65.51%, whereas when not using the SMOTE technique the Support Vector Machine algorithm has the highest performance. better than other algorithm models with an accuracy value on the data not 72.41%.

Every year more than 2 million Americans die from heart disease which is the number one killer in the world. The results of the Sample Registration System (SRS) survey show that heart disease is the highest cause of death at all ages after stroke, which is 12.9%. The method used in this study uses the Naïve Bayes algorithm. The purpose of this study is to determine if anyone with heart disease has a stroke. From the research results obtained by splitting data using 80:20 to get a prediction accuracy rate of 83% for heart disease prediction cases. In the trial results using the label test data obtained, namely no stroke.