Real-Time Data Processing Techniques for a Scalable Spatial and Temporal Dimension Reduction

Abstract

Wireless sensor networks (WSN) often generate data with high frequency per unit time. Values from sensor measurements are often redundant and provide little or no information. Methods of dimension reduction can be applied and thus only data of interest can be preserved for the purpose of effective analysis of wireless sensor networks' data. The authors show experimentally that it is possible to build more successful predictive models by reducing dimensions and discuss the potential advantages of reducing the spatial and temporal dimensions of sensor measurements in different applications. Authors present the implementation and analysis of an efficient distributed system that enables search, ranking, indexing, machine learning analysis and visualization of data from WSNs, processed in real-time.