重点研究了异常入侵检测系统模型。针对现有模型中存在的对训练数据要求高、误报率高等问题,提出了一种基于最小差异度聚类的入侵检测方法。该方法将区间标量、序数变量、二元变量标称变量类型的属性映射到区间[0,1]上,计算每个数据对象之间以及与各个簇的差异度,很好地解决了异常入侵。 Focus on the anomaly detection system model. Aiming at the problems of high training data and high false positive rate existing in existing models, this paper proposes an intrusion detection method based on minimum difference clustering. The method maps the properties of interval scalar, ordinal variable and binary variable nominal variable types to interval [0, 1], calculates the difference degree between each data object and each cluster, and solves the problem of abnormal invasion .