基于神经网络集成的入侵检测研究

Intrusion detection based on neural network ensemble

提出一种新的基于神经网络集成的入侵检测方法。首先通过有区别地对待不同的训练数据训练神经网络,提高对小类别入侵的检测能力并防止网络训练中的退化现象;然后利用一种新的改进遗传算法优化集成网络的权,提高系统整体性能。理论和实验表明该方法具有较好的检测能力。

This paper put forward a new intrusion detection method based on neural network ensemble. First, we trained neural network by using different methods to deal with different data, to elevate the ability of detecting small intrusion and prevent neural network degeneration during training; then, the improved Genetic Algorithm was used to optimize neural network ensemble weight. Theory and experiment show that this method has better ability in intrusion detection.