A probability-based approximate algorithm for anomaly detection in WSN

Anomaly detection is always a challenge for wireless sensor network (WSN). In some applications, researchers compare the number of abnormal nodes with the pre-provided threshold to detect the anomaly. But, it involves the whole abnormal nodes in WSN and consumes lots of resources. In this paper, we proposed a probability-based algorithm for anomaly detection which reduces involved nodes and consumes fewer resources. The experimental results demonstrate that, the algorithm can satisfies demand of query and is effective to reduce the cost of communication band and power energy by choosing appropriate parameters.