Prediction of Water Usage for Advanced Metering Infrastructure Network with Intelligent Water Meters

Abstract

In this article, we present the solution of a prediction algorithm, applied to real world Advanced Network Infrastructure, consisting of intelligent water meters. In the first step of the algorithm time series are passed to outlier detection in order to remove possible disturbing values. The prediction process is carried out with the use of four types of machine learning and deep learning algorithms. The proposed solution based on real world univariate time series taken from multi-family houses is evaluated. The experimental results confirm that the presented solutions are both efficient and flexible.