Network Intrusion Detection Through Classification Methods and Machine Learning Techniques
Dimitrios Simeonidis, Pavel Petrov, Jordan Jordanov · 2023
The field of network attacks is rapidly evolving, which makes it particularly difficult to grasp a universal picture of the problem. Intrusion detection is one of the most important tools for securing and protecting computer systems from malicious attackers by enabling detection of attacks and reducing their impact. In this research, we study the problem of cyberattacks and approach the possibility of securing computer and network communications by proposing intrusion detection approaches based on classification and machine learning mechanisms. Considering that intrusion detection is not simply a matter of predicting the most likely classification (attack-“normal” behavior), since different types of errors incur different costs, we propose the implementation of a cost-sensitive approach to intrusion detection and apply it to some well-known and efficient classification algorithms for both wired and wireless networks. The research method used in the study employs a research design focused on evaluating the effectiveness of intrusion detection systems using classification algorithms. Statistical analysis could be applied as necessary to analyze the data, and the research is designed to be reproducible for future studies.