ST-MLR: A Spatio-temporal Multiple Linear Regression Missing Data Reconstruction Approach for Improving WSN Data Reliability

Missing data is one of the unavoidable issues in Wireless Sensor Networks (WSNs) due to various reasons, including communication failure, unreliable communication links, unexpected damage, etc. WSNs are the base of many critical and non-critical applications, such as nuclear applications, medical applications, weather forecasting, etc. Therefore missing data reconstruction before their application or further analysis plays a vital role in data reliability. This paper proposed a missing data reconstruction approach based on the Multiple Linear Regression model (MLR) using Spatio-temporal correlation. The experimental results reveal that the proposed approach is effective and efficient in reconstructing missing data of different scales.