第一级认知系统在基于专家系统入侵检测中的应用研究

随着恶意入侵计算机现象的日益严重,准确检测入侵的需求应运而生。虽然基于专家系统的入侵检测方法检测准确度很高,但专家知识的获取和规则的动态更新是两大难点。该文提出将第一级认知系统(CS-1)应用于基于专家系统的入侵检测。CS-1采用桶队算法解决了规则评价和规则竞争问题,利用遗传算法不仅去除了降低检测效率的规则,而且取代由遗传算法生成的更强的规则子集。文中对专家系统规则到CS-1分类器的转换及分类器的检测方法进行了实例说明,实验结果体现了该方法的有效性和检测的高效性。

An Application Research of Cognitive System Level One in Expert System-Based Intrusion Detection

As malicious intrusions into computer system have become a growing problem,the need for accurately detect-ing these intrusions has risen.Although the detecting accuracy of Expert System-Based intrusion detection method is quite high,it is difficult to obtain expert knowledge and automatically update expert rules.This paper proposed CS-1(Cognitive System Level One)which is applied to Expert System-Based intrusion detection.The problem of evaluation and competition of rules has been solved by bucket brigade algorithm.The rules that have affected the efficiency of in-trusion detection will not only be deleted,but also be replaced by better rules that have been generated by genetic al-gorithm.An example to explain the conversion from the rules of Expert System to classifiers in CS-1and the detecting method of classifiers is also presented.The experiments show that the proposed method is practical and efficient.