基于BIRCH的入侵检测算法

针对已有的无线入侵检测方法训练时间长和检测精度低的问题,提出一种基于调整后的BIRCH——MBIRCH算法的无线Mesh网络入侵检测算法。该算法首先一次性扫描数据集获得CF（聚类特征）,然后自底向上地计算不同层次的聚类有效指标,主要是考虑数据集的几何结构,即通过度量簇内数据点分布的紧凑度以及簇间的相似度,并保持二者之间的平衡,根据此指标确定CF树的簇结点,直到得到最佳聚类结果,将最佳聚类结果作为训练样本指定判别函数,对网络数据定位。实验结果表明,该算法不仅明显减少样本训练时间,同时提高了算法检测精度,符合无线Mesh网络的入侵检测需要。

Intrusion Detection Algorithm Based on BIRCH

For the traditional IDS for wireless network is long in training time and low in detection accuracy,a WMN intrusion detection algorithm based on BIRCH is proposed.The method first obtains CF(clustering feature) by once only scanning the dataset,and then computes the hierarchical validity index,principally about geometrical structure of the dataset.Compactness in set and similarity between sets is measured and used to keep balance.The hierarchical validity index is used to confirm the positon of cluster node...