A benchmark for spatial and temporal correlation based data prediction in wireless sensor networks

Missing data is an inevitable problem in wireless sensor network and the way missing values are handled can significantly affect the analysis results involving such data. To address data missing issues, spatial correlation and temporal correlation modeling can be applied. This paper aims at reviewing some popular spatial and temporal correlation based methods. The proposed review includes a critical overview through a summary of pros and cons of these methods and a comparison between them based on simulation results. To our best knowledge, there is no such comparative benchmarking study in the current literature.