PERBANDINGAN 3 ALGORITMA KLASIFIKASI DATA MINING DALAM PRO-KONTRA BAHAYA ROKOK ELEKTRIK

Abstract

Many studies and facts exist in society that smoking is bad for health. This causes smokers to look for alternatives from traditional cigarettes to e-cigarettes (vaping) to help quit smoking slowly. However, recently the use of e-cigarettes has become its own pros and cons in society because of the acceptance of information from the community itself. This study tries to compare the use of 3 data mining classification methods, namely Decision Tree, Naïve Bayes and Logistic Regression to get the most accurate algorithm and find out the results of the classification of dangers or not using e-cigarettes based on the views of the community itself using 4 comparison factors, namely how often. see how people use/offer/advertise/promote e-cigarettes, how often they see shows about the dangers of e-cigarettes, have close friends who smoke vaping, and have family (father/mother/siblings) who smoke. From the results of data processing and testing by measuring the performance of the three algorithms using the confusion matrix procedure, operator cross validation and the ROC curve, the decision tree algorithm produces the highest level of accuracy value of 81.00% and based on the decision tree algorithm graph it can also be seen that you have Close friends who smoke e-cigarettes are the main factors that people think vaping is dangerous.