Comprehensive Performance Evaluation Of Network Intrusion System Using Machine Learning Approach

Abstract

Over the last three decades, network devices are increasing due to technology like the Internet of Things (IoT) and Bring Your Own Device (BYOD). These rapidly increasing devices open many venues for network attacks whereas modern attacks are more sophisticated and complex to detect. To detect these attacks efficiently, we have used recently available dataset UNSW-NB15. UNSW-NB15 is developed according to the modern flow of network traffic with 49 features including 9 types of network attacks. To analyze the traffic pattern for the intrusion detection system(IDS), we have used multiple classifiers to test the accuracy. From the dataset UNSWNB15, we have used medium and strong correlated features. All the results from different classifiers are compared. Prominent results are achieved by ensemble bagged tree which classifies normal and individual attacks with an accuracy of 79%.