基于隐马尔可夫模型的复合攻击预测方法

复合攻击成为网络攻击的主要形式之一，入侵检测系统仅能检测到攻击，但不能预测攻击。该文分析了传统的攻击预测方法的不足，提出一种基于隐马尔可夫模型的攻击预测方法，该方法使用隐马尔可夫模型中的Forward算法和Viterbi算法识别攻击者的攻击意图并预测下一步可能的攻击。通过实验验证了该方法的有效性。

Approach to Forecasting Multi-step Attack Based on HMM

The multi-step attack is one of the primary forms of the current attack. Intrusion detection system only detects attack, but cannt forecast next attack. This paper presents an approach to forecasting attack based on HMM after analyzing the defaults of some approaches to forecasting attack. To recognize the attacker’s attack intention and forecast next possible attack using the Forward and Viterbi algorithm. Experimental results prove the validity of the approach.