A multiple linear regression data predicting method using correlation analysis for wireless sensor networks

Abstract

For there are too many drawbacks like excessive input variables, high computational complexity and low efficiency in evaluation methods of missing data for wireless sensor network, A Multiple-Regression evaluation method based on correlation analysis was proposed in this paper. First, the sensor data of the wireless sensor networks was correlatively analyzed, and most correlation sensor data was explored. Then the sensor data was used as input of the multiple linear Regression model and evaluation method. In the experimental stage, the sensor temperature data of actual wireless sensor network has been used to test this method. Experiment results show that the scheme is efficient with low prediction error, thus it owns practical value and can be used to evaluate the missing data in wireless sensor networks.