OPTIMASI ALGORITMA KLASIFIKASI DECISION TREE (CART) DENGAN METODE BAGGING UNTUK DETEKSI WEBSITE PHISHING

Abstract

Advances in science and technology (IPTEK), especially information technology, make it easier to support community activities, but at the same time make various devices vulnerable to exploitation by cybercriminals [1]. One such activity is stealing data from Internet users through fake sites (also known as phishing sites) designed to look like the real thing. Phishing websites pose a serious threat to online information security and require an effective approach to detect and prevent them. To combat the proliferation of phishing websites in cyberspace, a classification is needed to predict which websites will be classified as phishing websites using the Decision Tree Classification Algorithm (CART) [2]. To improve the performance of the Decision Tree Classification (CART) algorithm and achieve better optimal accuracy, optimization using the bagging method is needed. A bagging technique that combines the results of several decision tree models is applied to improve the performance and reliability of the CART algorithm in detecting phishing websites [3]. In this research, we collected a dataset containing various characteristics related to the characteristics of phishing websites. The data is then processed and divided into subsets for model training and testing. The aim of this research is to optimize the decision tree classification algorithm (CART) by applying bagging techniques in the context of phishing website detection. Based on test results, applying the Decision Tree Classification Algorithm (CART) to classify phishing websites produces an accuracy of 96.61%, and when combined with bagging techniques, the accuracy increases by 1.13% to 97.74%. This experiment shows that optimization can improve the prediction accuracy of phishing websites by combining the Decision Tree Algorithm (CART) with bagging techniques.

Keywords: Phishing Websites, Classification, Prediction, Algorithms, Decision Trees.