Finding Group-Based Skyline over a Data Stream in the Sensor Network

Along with the application of the sensor network, there will be large amount of dynamic data coming from sensors. How to dig the useful information from such data is significant. Skyline query is aiming to identify the interesting points from a large dataset. The group-based skyline query is to find the outstanding Pareto Optimal groups which cannot be g-dominated by any other groups with the group same size. However, the existing algorithms of group-based skyline (G-Skyline) focus on the static data set, how to conduct advanced research on data stream remains an open problem at large. In this paper, we propose the group-based skyline query over the data stream. In order to compute G-Skyline efficiently, we present a sharing strategy, and based on which we propose two algorithms to efficiently compute the G-Skyline over the data stream: the point-arriving algorithm and the point-expiring algorithm. In our experiments, three synthetic data sets are used to test our algorithms; the experiments results show that our algorithms perform efficiently over a data stream.