基于时空相关性的传感器网络离群点检测算法  

SPATIOTEMPORAL CORRELATION-BASED OUTLIER DETECTION ALGORITHM IN WIRELESS SENSOR NETWORKS

作　　者：林锋[1] 张红[1] 

机构地区：[1]浙江经贸职业技术学院浙江杭州310018

出　　处：《计算机应用与软件》2013年第6期114-115,221,共3页Computer Applications and Software

基　　金：浙江省教育厅科研项目(Y201016416)

摘　　要：离群点检测是传感器网络中的一个重要问题,它主要包括事件检测、异常检测。提出一种基于时空相关性的、轻量级的局部离群点检测算法,它可利用节点的计算能力,降低节点之间的通信成本;同时,还能区分出离群点的类型。Outlier detection is an important issue in sensor networks, which mainly comprises event detection and anomaly detection. The detection algorithm proposed in the paper is a lightweight one based on spatiotemporal correlation for local outliers. It can utilise the computational capacity of the sensors adequately and reduce the communication cost between the nodes. It can also differentiate outlier types.

关 键 词：离群点 传感器网络 时空相关性 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]