自适应多趟聚类在检测无线传感器网络安全中的应用

摘要

On the basis of KSummary algorithm,introduce hierarchical and density clustering method,propose adaptive multi-trip clustering method. In turn obtain number of clustering k,clustering initial centers and the final clustering. The algorithms is used in wireless sensor networks( WSNs)data,and can find out outliers of datas very well,thereby find existing security risks of sensor nodes. Experimental results and analysis show that the algorithm can not only obtain stable and convergent clustering results,but also can find out outliers very well.%在KSummary算法的基础上，引入层次和密度聚类方法，提出自适应多趟聚类方法。依次获得聚类个数k，聚类初始中心和最终聚类。将算法应用于无线传感器网络数据中，可以很好地发现数据中的离群点，从而找到传感器节点安全上存在的隐患。实验结果和分析表明：此算法不但可获得稳定、收敛的聚类结果，还能很好地发现离群点。