USING MACHINE LEARNING AND DEEP LEARNING TO IMPROVE ANOMALY ATTACK
Thanh-Huy Nguyen, Van‐Dinh Nguyen, Thanh-Huy Nguyen, Van Son Nguyen, Thai Thanh Tung, Tran Tien Dung, Le Thi Thanh Thuy, Nguyen Liem Hieu, Nguyen Minh Dung, Nguyen Van Ba · Journal of Southwest Jiaotong University · 2023
The growth of the Internet and the increasing frequency and sophistication of cyberattacks have made effective intrusion detection systems a critical component of network security. NIDS is always a necessary solution for monitoring and detecting the most common attack types from a hacker to an organization. This paper proposes an effective approach for detecting cyber threats in an intrusion detection system using the deep learning algorithm dataset. We research many perspectives such as decreasing the false positive rate of detection, increasing detection time, combining various deep learning algorithms, and mixing multiple types of datasets. In addition, we propose a practical solution for this research that can be used in a production environment when deployed and integrated comprehensively with traditional NIDS solutions to generate near real-time alerts to help efficiently protect enterprise networks. The goal of this research is to build NIDS that combines deep learning algorithms and diverse data types to improve the performance of hacker attack detection and enhance network security. This research proposes a unique way to efficiently combine machine learning/deep learning with new data sets in NIDS. The breakthrough of this approach lies in its ability to identify complex relationships between data and deep learning models, allowing optimized attack detection and increased flexibility in handling new threats. By leveraging the potential of diverse data sets, this approach promises to shape an important step forward in enhancing the performance and efficiency of NIDS systems in real-world environments. Keywords: Cybersecurity, Intrusion Detection System, Machine Learning, Deep Learning, Anomaly Detection DOI： https://doi.org/10.35741/issn.0258-2724.58.4.40