摘要
To solve the problem of the naive bayes algorithm in intrusion detection inside the high data dependency and decision tree to data "broken" problem, this paper combines the advantages of decision tree section and the advantages of simple bayesian evidence more fusion, based on the bayesian process service model for prediction of tree algorithm, and to integrate the bagging learning method is used to improve the bayesian tree, the experimental results show that the model can effectively detect abnormal host, and the time complexity of the algorithm is relatively low, suitable for online detection.
- 出版日期2015
- 单位西安近代化学研究所; 中国人民武装警察部队工程大学