Research on the Large-scale Network Intrusion Mode based on Principal Component Analysis and Drop Quality Sampling

Abstract

In this paper, we conduct research on the large-scale network intrusion mode based on the principal component analysis and drop quality sampling. With the growing of network security issues, invasion detection becomes the study hotspot. There are two main types of that invasion detection technology, the first is that misuse detection and the anomaly detection. Misuse detection can more accurately detect attacks, but high non-response rates, anomaly detection could detect the unknown attacks, but higher rate of false positives. Network invasion detection problem is summed up in the network data flow of discriminant problem, namely the judgment of network data flow is normal or malicious and in this sense here invasion detection problem can be understood as a pattern recognition problem. Our research integrates the PCA and sampling technique to propose the new idea on the IDS that is innovative and will promote the development of the corresponding techniques.