Statistical Intrusion Detection in Data Networks

The growth of application of electronic commerce relying on Internet and the network attack have become more common and sophisticated. The network systems demand more elaborate protections measures to ensure safely operation, and continue provide critical services. These measures include meth- ods of detection and response to intrusion attempts in real time. In this work we propose statistical models and multivariate classifiers to detect anomalous traffic shape and the use of neural network classifier. Numerical results are shown for the simples case of denial of service attack.