Exploring the Impact of PCA Variants on Intrusion Detection System Performance

Abstract

Intrusion detection systems (IDS) play a critical role in safeguarding network security by identifying malicious activities within network traffic. However, the effectiveness of an IDS hinges on its ability to extract relevant features from the vast amount of data it collects. This study investigates the impact of different feature extraction methods on the performance of IDS. We compare the performance of various feature extraction techniques on two widely used intrusion detection datasets: KDD Cup 99 and NSL-KDD. By evaluating these techniques on both datasets, we aim to gain insights into the generalizability and robustness of each method across different dataset characteristics. The study compares the performance of these methods using standard metrics like detection rate, F-measure and FPR for intrusion detection.