Big Data pre-processing techniques within thewireless sensors networks

Abstract

The recent advances in sensors and communications technologies have emerged the interaction between physical resources and the need for sufficient storage volumes for keeping the continuously generated data. These storage volumes are one of the components of the Big Data to be used in future prediction processes in a broad range of fields. Usually, these data are not ready for analysis as they are incomplete or redundant. Therefore one of the current challenge related to the Big Data is how to save relevant data and discard noisy and redundant data. On the other hand, Wireless Sensor Networks (WSNs) (as a source of Big Data) use a number of techniques that significantly reduce the required data transmissions ratio. These techniques not only improve the operational lifetime of these networks but also raise the level of the refinement at the Big Data side. This article gives an overview and classifications of the data reduction and compression techniques proposed to do data pre-processing in-networks (i.e. in-WSNs). It compares and discusses which of these techniques would be adopted or modified to enhance the functionality of the WSNs while minimizing any further pre-processing at the Big Data side, thus reducing the computational and storage cost at the Big Data side.