This research focuses on method for data location of abnormal node based on cascade-trapped wave with self-adaption.Firstly,we started with degree of approximation of sampling node data in sliding window and judged whether node data in window is abnormal integrated with cluster theory,then adjusted judgment of self-judgment stage of network node according to neighbor information.The research designed structure of lattice filter with second order and used multiple fixed cascade of wave trap to suppress interference component of data attribute of abnormal node.Finally,the research quested optimization characteristic solution integrated with matching projection theory and found out all matched feature dot pairs.Thus,we achieved the data location.Simulation proves that the method can improve accuracy of data location effectively and has preferable anti-disturbance performance.%对大数据网络环境下异常节点数据的定位研究,可以有效降低网络空间存在的安全威胁和存储开销.对异常节点数据的定位,需要结合匹配投影理论寻求优化特征解,找出所有匹配的特征点对.传统方法将节点动态感知数据聚合成变宽的直方图,来准确定位节点异常数据,但忽略了求取节点的匹配特征点对,导致定位精度较低.提出基于自适应级联陷波的异常节点数据定位方法.从滑动窗口内采样节点数据的近似度出发,结合聚类理论思想判断窗口内的节点数据是否异常,依据邻居信息将网络节点自我判断阶段的判决进行调整,设计二阶格形滤波器结构,用多个固定陷波器级联抑制异常节点数据属性干扰成份,结合匹配投影理论寻求优化特征解,找出所有匹配的特征点对,从而实现异常节点数据定位.仿真证明,所提方法能够有效提升异常节点数据定位精度,且具有较好的抗干扰性能.
展开▼
