Intrusion Detection System Based on Machine Learning
Bocheng Liu, Zhiqin Huang, Zeguo Zhu · Proceedings of the 7th International Conference on Cyber Security and Information Engineering · 2022
With the continuous development of the network, various Internet Plus are constantly derived, but more and more network threats also arise spontaneously. Intrusion Detection Systems (IDS) have become an important part of defending against malicious network attacks due to their ability to take proactive defenses. This paper compares four malicious traffic detection algorithms based on machine learning: through feature extraction and normalization of the data, and then brought into the model for training, comparison and improvement. Finally, an IDS based on random forest algorithm is designed to identify malicious http requests and give network administrators a better feedback. The accuracy rate can reach 0.95, the recall rate can reach 0.96, and the f1 value can reach 0.95.