摘要： 针对目前入侵检测中存在的误检率高、对新的入侵方法不敏感等问题，提出了一种基于神经网络集成的入侵检测方法。使用负相关法训练神经网络集成，采用tf×idf的系统调用编码方式作为输入。实验结果表明，与单神经网络方法相比，神经网络集成弥补了神经网络方法在检测数据上的不足，在保证较高的入侵检测率的前提下，保持了较低的误检率。

关键词: 入侵检测, 神经网络集成, 负相关学习

Abstract: For the problem of high false-positive rate and not sensitive to new intrusion method existed in the intrusion detection, a new intrusion detection method based on neural network ensemble is proposed. It trains the neural network ensemble with negative correlation learning method, uses tf×idf(term frequency ×inverse document frequency)system calls encoded mode as the input. The experimental results indicate that, compared with ANN, neural network ensemble improv the performance of ANN in the data analysis, keeps high detection rate and low false-positive rate.

Key words: Intrusion detection, Neural network ensemble, Negative correlation learning

中图分类号: 

• TP391