基于模糊数据挖掘与遗传算法的异常检测方法

建立合适的隶属度函数是入侵检测中应用模糊数据挖掘所面临的一个难点。针对这一问题，提出了在异常检测中运用遗传算法对隶属度函数的参数进行优化的方法。将隶属度函数的参数组合成有序的参数集并编码为遗传个体，在个体的遗传进化中嵌入模糊数据挖掘，可以搜索到最佳的参数集。采用这一参数集，能够在实时检测中最大限度地将系统正常状态与异常状态区分开来，提高异常检测的准确性。最后，对网络流量的异常检测实验验证了这一方法的可行性。

Anomaly detection approach based on fuzzy data mining and genetic algorithm

Defining appropriate membership functions is a difficult task in fuzzy data mining to detect intrusions. To solve the problem, an approach that applies genetic algorithm to optimize parameters of membership functions in anomaly detection was presented. Parameters of membership functions were arranged into a sequential parameter-set coded to an individual. An optimal parameter-set could be derived by embedding fuzzy data mining in the process of evolution of individual. With the parameter-set in anomaly detection, normal state of protected system could be differentiated from anomalous state in the most extent, and the veracity of anomaly detection was improved greatly. Experiments on anomaly detection to network tratffic prove the feasibility of the approach.