Online Algorithms for Mining Inter-stream Associations from Large Sensor Networks

We study the problem of mining frequent value sets from a large sensor network. We discuss how sensor stream data could be represented that facilitates efficient online mining and propose the interval-list representation. Based on Lossy Counting, we propose ILB, an interval-list-based online mining algorithm for discovering frequent sensor value sets. Through extensive experiments, we compare the performance of ILB against an application of Lossy Counting (LC) using a weighted transformation method. Results show that ILB outperforms LC significantly for large sensor networks.