Detection of DoS Attack using AdaBoost Algorithm on IoT System
Salman Rachmadi, Satria Mandala, Dita Oktaria
Abstract
Internet of Things (IoT) is a networking concept where an object can transmit data over the internet without any human interaction. The object is usually a sensor with a communication device connected to the Internet. The popularity of IoT is increasing with the advent of 5G technology. However, the threats to the system are also getting more intense. One of the serious threats to IoT systems is known as denial of service (DoS) attack, which usually target broker services on that system. Several researches have been performed to overcome this DoS attack. However, the results appear to be ineffective. It can be seen that the accuracy of the DoS detection systems are still low. This study aims to provide a solution to the above problems by proposing an Intrusion Detection System based on Artificial Intelligence (AI, AdaBoost) for IoT system. The method used in this study is supervised learning which measures the accuracy of predictions in detecting DoS on IoT network data. The experiments have been carried out on 130223 DoS attack data and 130284 normal data. The detection accuracy of the DoS detection is 95.84 % and the F1-Score is 95.72 %. Recall and precision have achieved 93.28% and 98.29%, respectively.