Application of Bayesian Tree Algorithm to Anomaly Intrusion Detection
Jing Xu · Journal of Jiangsu University of Science and Technology · 2008
The anomaly intrusion detection algorithm on Windows platform is studied.With the Windows Native API sequence,a host process rule and the corresponding probability distribution are presented as a normal model based on the Bayesian tree algorithm.With the test sequence,the probability distribution of the nth native API from the preceding N-1 native API is predicted by using the U-test method for the anomaly intrusion detection algorithm.Moreover,the AdaBoost-M1 algorithm is used to construct a series of probability distribution sequences based on the Bayesian tree algorithm.At last,a final probability distribution based on the boosted combination of the probability distribution sequences is formed,which can be applied to the intrusion detection.Experiment results proved a good prediction performance with this model.