KOMPARASI ALGORITMA NAÏVE BAYES DAN ALGORITMA C4.5 DALAM KLASIFIKASI PELANGGAN PRODUK INDIHOME

Abstract

Abstract Rapid technological advances encourage every company, including PT. Telkom Indonesia for its product namely IndiHome, to keep abreast of technological developments and continue to improve its ability to manage data and information that is more secure, accurate, and efficient, while the problems experienced are difficulties in managing data properly to classify potential and non-potential customers. In comparing this research using a classification model is the grouping of objects into certain classes based on the group which is usually called a class. One of the classification methods that is often used is the Naive Bayes method and C4.5 using Rapidminer tools. The stages of this research use the Cross-Industry Standard Process for Data Mining (CRISP-DM) flow which has stages of Business Understanding, Data Understanding, Data Preparation, Modeling, Evaluation, and Deployment. The purpose of this study is to obtain a pattern of knowledge on the classification of IndiHome customers who have potential and non-potential customers so that they can be used as a source of decision making by PT Telkom Indonesia. The dataset used is 1043 records with the highest accuracy results obtained in the Naive Bayes classification getting an accuracy value of 99.71% with a value of classification error 0.29%, precision 100%, recall 99.35%, and an AUC value of 0.999 in the Excellent Classification category. While the classification algorithm C4.5 obtained a lower accuracy value of 89.94% with a value of classification error 10.06%, precision 89.30%, recall 87.96%, and an AUC value of 0.899 into the Good Classification category.   Keyword : Naive Bayes Classifier (NBC), C4.5, IndiHome, CRISP-DM