Water is an essential and crucial element of each of the living beings. The urban water systems cater to the water necessities of the communities. Hence urban water management is a very critical task. Water demand modeling and prediction is one of the techniques that elevates resource allocation, supply cost reduction, fatalities supply network reduction, and so on. We are proposing a water demand prediction platform with the existing dependent features and a deep neural network model to incorporate the time series, weather, and demand to predict the short-term water demand effectively. Since the IoT sensors are the major data sources for this prediction model we are proposing a multivariate data imputation model, multi-variate Gaussian- based GAIN (Generative adversarial imputation Net).
Name of students and staff from Amrita : Nibi K V, Research Associate Amrita Center for Wireless Network & Application
International Collaborators : Dragan Savic FREng, Professor University of Exeter & Univ. of Belgrade; KWR Water Research
