Prediction for Diagnosing Liver Disease in Patients using KNN and Naïve Bayes Algorithms

Abstract

There is a lot of data on patients who undergo medication or medical examinations at the hospital and this is information that must be extracted so that it can provide information for future improvement conditions, meaning that past data can be used as a prediction basis for liver disease in patients. This is very beneficial for medical personnel and also for patients if they experience symptoms that match the symptoms felt by a patient. This project uses machine learning because it involves big data and past data is used to predict future data. Referring to previous research in reference that the results of the evaluation are varied. So in this study, The proposed strategy to performance optimization is carried out based on training data and variables that affect the model. Based on the results of calculations and analysis, it was found that the performance evaluation values were area under the curve(AUC) for naïve Bayes algorithm is 72.5% and k-nearest neighbour (KNN) of 63.19%.