A novel association rule mining mechanism in wireless sensor networks

Abstract

Sensor association rules have been found to be very useful for improving the quality of service, energy conservation, resource management, etc. in wireless sensor networks (WSN). Most of the association rule mining algorithms for sensor networks require the behavioral data, which describes the sensor activities, to be sent to the central node (sink node) by the sensors. Then, the sink node creates the sensor database and applies different algorithms to find association rules from the sensor database. In this paper, we propose an in-network mechanism to find frequent sensor patterns in the sensors themselves. So, the sensors send only the frequent senor patterns to the sink, not the sensor activity sets.