Implementasi Algoritma Naïve Bayes Pada Data Set Hepatitis Menggunakan Rapid Miner

Abstract

One of the diseases anticipated by doctors is hepatitis. This is because if the patient is not detected from the beginning having hepatitis, then the disease will develop into liver cancer. It can be seen that cancer is one of the deadliest diseases in the world that there are no drugs used for healing. By utilizing this increasingly developing science, researchers try to predict or classify whether a patient has suffered from hepatitis sickness based on the results of tests that have been undertaken before. One data mining technique can be used to predict Hepatitis and the method used is Naive Bayes. The data used is sourced from the UCI Repository with the web address https://archive.ics.uci.edu/ml/datasets/Hepatitis . The amount of data available is 155 data with 123 patients with Life hepatitis and 32 patients with Die hepatitis. The attributes contained in this hepatitis dataset are: Age, Sex, Steroids, Antivirals, Fatigue, Malaise, Anorexia, Big Liver, Liver Firm, Spleen Palpable, Spiders, Ascites, Varices, Bilirubin, Alk Phosphate, Sgot, Albumin, Protime, Histology , and Class (predictive result attribute). From the results of the research that has been done, it can be concluded that the Naive Bayes method includes an accurate algorithm to predict because the results of accuracy using Rapid Miner show more than 50% which is equal to 76.77%. With the highest Precision Class results of 98.88% for "Life" predictions, and Class Recall of 96.88% for "Die" Predictions.