Spatio-temporal correlation based outlier detection algorithm in sensor network

Sensor nodes are prone to produce faulty sample data because of poor quality and harsh environment. In this paper we put forward STCOD algorithm which uses spatial and temporal correlations between sensor nodes to distinguish faulty sample data from outlier data. STCOD algorithm includes three child algorithms which are outlier self-detection algorithm, neighbor-voting algorithm and outlier confirming algorithm. Experiments show that STCOD outperforms majority voting algorithm.