Comparative Analysis of Machine Learning Models for Data Traffic Anomaly Detection Systems in Intrusion Detection Systems
Luthfi Hakim Panuntun, Nabilah Nawang Amaranggani, Syafrizal Ananta Adhitya, Amiruddin Amiruddin, Sri Rosdiana · 2023
Cyberattacks can happen at any time and have many different goals. It might take the shape of ransomware, DOS attacks, or even malicious software. Its existence is undoubtedly a reason for concern since a continuous internet network might result in data flooding and a decline in the network's performance. In this paper, we try to develop an application of machine learning to analyze data traffic in the network. As we know, in a data traffic network, there are various types of data traffic in units of time. Analyzing each incoming network is very time-consuming, which makes it inefficient and ineffective so we try to develop an automated machine that can detect data traffic, or in other terms an Intrusion Detection System. By analyzing the types of attack variables, the KDD 1999 dataset provided one of the key components for constructing an IDS system. There are seven common types of supervised machine learning models. The machine learning model is designed to investigate the various types of attacks contained in the KDD 1999 dataset. The LGBM model is the most accurate of the seven models.