ARELLANO, Alfonso    PENA, Daniela. Linear regression models for predicting drinking water consumption. []. , 3, 1, pp.27-36. ISSN 2631-2654.  https://doi.org/10.37135/ns.01.05.03.

This research provides two predictor models of drinking water consumption for the residential sector. They would serve the designers to define the endowments required by a population. Variables affecting drinking water consumption are grouped into sociodemographic, socioeconomic, water management and quality, and climatological. Multiple linear regressions are performed. Variable coefficients are obtained to define two mathematical models. A model calculates the six-monthly per capita consumption (CPC/est.s) with information about each socioeconomic level (R2 adjusted=80.87%). It requires 19 variables. The second model estimates weighted monthly per capita consumption (CPC/p.m) (R2 adjusted=38.88%). It requires six variables. The water management and quality and demography variables are significant in the six-monthly per capita consumption. The climatological variables humidity and maximum temperature have predominant incidences in weighted monthly per capita consumption. The two models can predict water consumption to ensure a rational resource endowment in distribution systems. Because of the variables' dynamic nature, the information should be updated continuously to ensure the results' sustainability.

: Drinking water; mathematical models; predictors.

        · |     · |     · ( pdf )