Parzen-window network intrusion detectors

摘要: Network intrusion detection is the problem of detecting anomalous network connections caused by intrusive activities. Many systems proposed before use both normal and data to build their classifiers. However, are usually scarce difficult collect. We propose solve this using a novelty approach. In particular, we take nonparametric density estimation approach based on Parzen-window estimators with Gaussian kernels an system only. To facilitate comparison, have tested our KDD Cup 1999 dataset. Our compares favorably winner which ensemble decision trees bagged boosting, as uses no at all much less for training.