Modelos de aprendizagem de máquina para previsão da demanda de água da região metropolitana de Salvador, Bahia.

Abstract

Water is essential for human life and predicting its growing demand is a big challenge, the objective of this work is to propose a new hybrid model, SVR-ANN, to water demand forecast, where an adaptation of the methodology proposed by Zhang (2003) is used to decompose the time series of 10 reservoirs that supply the Metropolitan Region of Salvador (RMS). The data used is from historical consumption from January/2017 to February/2022, in liters per hour, obtained from the local supply, Bahian Water and Sanitation Company (EMBASA), and meteorological data obtained from the Brazilian National Meteorology Institute (INMET). You results demonstrated the feasibility of using the proposed model, compared to other traditional models like Multilayer Perceptron (MLP), Support Vector Regression (SVR), Short Long Term Memory (LSTM) and Autoregressive and Integrated Moving Average (ARIMA).