基于数理统计及人工免疫的入侵检测模型

Intrusion detection system model based on mathematic statistic and immunlogical principle

分析了生物免疫系统的机制,介绍了基于免疫原理的入侵检测系统,针对目前计算机免疫机制和算法存在的不足,结合入侵检测机制,提出了运用HAMMING距离数理统计算法来改进特征码的模式匹配规则,并提出了一个包含基于神经网络的模式识别器的入侵检测模型。入侵检测系统根据免疫机制能够在网络攻击一旦发生时,动态地、自我学习性地产生高质量的核心检测元。实验结果表明,运用上面提出的算法及模型能够很好的提高系统的非线性和自适应性。

The detailed analyses have been made on the immunological principle. The intrusion detection systems based on immunological principle are introduced. According to the disadvantage of artificial immune theory and current IDS, the rule of the pattern matching about character codes by applying the mathematic statistic on the distance of HAMMING with immune theory is improved. Above that neural network algorithm is added into a new immune detector. It provides the capacity that gathering the high-quality core elements dynamically and self-studiedly as soon as network intrusions take place. The experiment has proved that though the improvement the system has been effectively accelerated. The non-linearity and adaptively of model has also been advanced.