Network Intrusion Detector Based On Isolation … Forest Algorithm
S Snehaa, G. R. Shwetha, B. Priya · 2022 1st International Conference on Computational Science and Technology (ICCST) · 2022
The internet serves as the global marketplace. A computer network will undoubtedly be required for every organization tobe successful. Connecting a company to the internet dramatically increases its reach, value, and efficiency. When linking a business to a network, however, security becomes a major problem because data is vulnerable to malevolent users. This is when having an “Intrusion detection system (IDS)” comes in handy. The ultimate aim of this project is to create a network intrusion detector based on the isolation forest algorithm: Machine Learning Unsupervised Learning, a model that can discriminate between intrusions, which are bad connections, and regular connections, which are good. Instead of profiling normal points, this model-based strategy specifically isolates anomalies. On the benchmark NSL-KDD dataset, the test has been performed. In comparison to several other intrusion detection technologies, the proposed technology obtains a high accuracy rate of 88 percent and above.