Output details

Clustering Distributed Time Series in Sensor Networks

This paper proposed a novel method for clustering sensor nodes in the network based on similarity of produced time series and proximity among densely deployed wireless sensor nodes. The novel clustering technique has been thoroughly tested on real datasets showing results proving it to be the best performing incremental and in-network clustering in wireless sensor networks for such experimental settings. The technique was developed in response to the need to have intelligent irrigation system preserving water, and was supported financially by the Australian government. The paper is published in ICDM, the world premier data mining conference.