Forecast model of water consumption for Naples

Traversal grey model for water consumption prediction

“…Table Ш shows that the MAE value and MAPE value of P(1)-GM(1,1) (year=1998) ， namely, the traditional GM(1,1), are 1.52×10 6 m 3 /a and 6.32% respectively, is larger than that of P(6)-GM(1,1) (year=2003), its MAE value and MAPE value are 0.78×10 6 m 3 /a and 3.57% respectively. It is confirmed that traditional GM (1,1) is not the best model in the n P(k)-GM (1,1).…”

“…It was calculated that the posterior error ratio value C of T-GM(1,1) is 0.0705<0.35,and the small error probability value P is 100%≥0.95, which shows T-GM (1,1) in accordance with first-order model, this forecasting model is credible. Table Ш shows that the MAE value and MAPE value of P(1)-GM(1,1) (year=1998) ， namely, the traditional GM(1,1), are 1.52×10 6 m 3 /a and 6.32% respectively, is larger than that of P(6)-GM(1,1) (year=2003), its MAE value and MAPE value are 0.78×10 6 m 3 /a and 3.57% respectively.…”

“…With the consequence of increased water consumption over the last few years and frequency of water shortages which have been reported in various areas, a closer examination of the dynamics of this evaluation is thus necessary [1], and scientific analysis and prediction is to satisfy demand, guarantee urban sustainable utilization of water resources [2]. Literature shows that other researchers have already carried out research on the GM(1,1) model which demonstrates the yearly water consumption time series [3][4][5].…”

System dynamics modeling for municipal water demand estimation in an urban region under uncertain economic impacts

“…X Molino et al (1996) A time evolution model of water consumption for prediction of short-term water demand using autoregressive moving average.…”

Técnicas de predicción a corto plazo de la demanda de agua. Aplicación al uso agrícola

“…La primera consiste en establecer modelos matemáticos basados en la correlación entre los datos de demanda y factores demográficos y ambientales (Maidment et al, 1985;Saporta y Muñoz, 1994;Rüfenatch y Guibentif, 1997), mientras que la segunda modela la relación entre datos presentes y pasados de la demanda (análisis estocástico de series temporales) (Coulbeck et al, 1985;Hartley y Powell, 1991;Jowitt y Xu, 1992;Shvartser et al, 1993;Saporta y Muñoz, 1994;Molino et al, 1996;Nel y Haarhoff, 1996). La primera metodología es poco usual, dado que recoger los datos referentes a factores considerados determinantes es igual o si cabe más complicado que recoger los propios datos de consumo, y la inclusión de estos factores se contempla de forma implícita a través de las observaciones de la demanda (Saporta y Muñoz, 1994).…”