基于免疫的网络入侵检测与风险预测模型

摘要

This paper presents a dynamic network intrusion detection and risk prediction model based on artificial immune thought, gives network attack detection process, antibody concentration calculation method and risk prediction process of Auto-regressive sliding average model based on time squence. Experimental results show that the model can quantitative analyze the current network risk and predict the future risk, and it can detect network risk real-time and has good prediction results. Compared with GM(1,1) model, it has more accurate results.%借鉴人工免疫思想,提出一种动态网络入侵检测与预测模型,给出网络攻击检测过程、抗体浓度计算方法及基于时间序列的自回归滑动平均模型的风险预测过程.实验结果表明,该模型可实时定量地分析网络当前的安全态势并对网络面临的风险做出预测,对于突变性网络风险预测效果优于GM(1,1)模型,且与实际风险状况较接近,具有较高预测精度.