一个基于改进遗传算法的RBF网络入侵检测模型

针对异常入侵检测中存在的误报率高的问题,文章提出了一种基于改进遗传算法的RBF网络入侵检测模型。采用数据挖掘方法建立聚簇规则集,用改进的遗传算法优化RBF网络,用已训练好的RBF网络对与聚簇规则集中不匹配的可疑行为进行检测,并能识别出具体的入侵类型。实验表明,文中提出的模型采用改进遗传算法的RBF神经网络,较基于BP神经网络的检测技术有更好的识别精度。

An intrusion detection model based on RBF neuron network of improved genetic algorithm

Aiming at the high rate of false alarm in anomalous intrusion detection, this paper puts forward an intrusion detection model based on RBF neuron network of improved genetic algorithm(GA). In this model, clustering rule set is established through data mining methods and RBF neuron network is optimized through improved GA. The suspicious behavior unmatched with anyone of the rules is detected through trained RBF neuron network, and the specific types of intrusion can be identified. Experiments show that the model using of RBF neuron network of improved GA, compared to BP neuron network, achieves better accuracy of identification.