Perbandingan Algoritma Decision Tree dan K-Nearest Neighbor untuk Klasifikasi Serangan Jaringan IoT

Abstract

As the number of uses of the Internet of Things continues to increase and expand. Security threats on IoT networks are also increasing. There are several techniques applied to overcome this security threat. One of them is a technique to classify an activity that is included in an attack or not along with the type of attack. Machine learning can be utilized for this classification process. Among the machine learning algorithms that can be used for this research are the Decision Tree and K-Nearest Neighbor algorithm approaches. This research aims to get the best classification results to detect the type of IoT network attack in both binary classification and multiclass classification. This research utilizes the Edge-IIoTset Cyber Security Dataset of IoT & IIoT. The results of the evaluation values obtained show that the performance of the Decision Tree algorithm is better than the KNN algorithm. With the difference in precision, recall, F1-score, and accuracy values are 0.15, 0.18, 0.17 and 0.08 in binary classification, respectively. While in multiclass classification, the difference value between the two algorithms is 0.26, 0.20, 0.22, and 0.23 respectively for precision, recall, F1-score, and accuracy