A Missing Data Imputation Algorithm in Wireless Sensor Network Based on Minimized Similarity Distortion

Abstract

This paper presents a novel wireless sensor network data imputation algorithm based on minimized similarity distortion (MSD). Firstly, the MSD algorithm considers attributes of the sensor datasets besides spatial and temporal to achieve complete dimensional data segmentations. It improves the problem of ignoring both the relationship of different attributes and the similar details in local data area. After that, it computes the distance between data units to get the k-nearest neighbors of the data units with missing values. For every missing value, MSD gives K preliminary predictive values with linear regression. Finally, MSD take the weighted K values as the final predictive values. Experimental results on real public wireless sensor data sets are provided to illustrate the efficiency and the robustness of the proposed algorithm.