Enhanced water demand analysis via symbolic approximation within an epidemiology-based forecasting framework

摘要: Epidemiology-based models have shown to successful adaptations deal with challenges coming from various areas of Engineering, such as those related energy use or asset management. This paper deals urban water demand, and data analysis is based on an Epidemiology tool-set herein developed. combination represents a novel framework in hydraulics. Specifically, reduction tools for time series analyses symbolic approximate (SAX) coding technique able simple versions sets are presented. Then, neural-network-based model that uses SAX-based knowledge-generation improve forecasting abilities. knowledge produced by identifying distribution district metered high similarity given target area sharing demand patterns the latter. The proposal has been tested databases Brazilian utility, providing key improving management hydraulic operation system. shows several benefits terms accuracy performance neural network demand.