本文通过分析现有入侵检测技术的优缺点,提出一种改进的聚类算法CLOPE将其应用 于网络入侵检测。该算法不需标准化数据,应用一种基于统计直方图的全局准则函数进行聚类,特别适用于大规模数据。在运用该算法时,还优化了收益值的计算,加快运行速度。实验证明,该算法不同于一般的聚类算法,能大大提高入侵检测的效率。 关键词: 全局准则函数,聚类算法,入侵检测 Abstract: This article through the analysis existing intrusion detection technology good and bad points, proposed one kind of improvement cluster algorithm CLOPE to apply to the network intrusion detection. This algorithm does not have to standardize the data, uses a kind of global criterion function based on the statistical histogram to depict clusters, specially is suitable for the large-scale data. When utilizes this algorithm, we optimized the income value computation, speeds up the running rate. The experiment proved that, this algorithm is different from the general cluster algorithm , it can improve and invade the efficiency measured greatly. Keywords:Global criterion function,Cluster alogorithm,Intrusion detection system