Assessing domestic factors determining water consumption in a semi-arid area (Sedrata City) using artificial neural networks and principal component analysis

Abstract

The growing demand for fresh water and its scarcity are the major problems encountered in semi-arid cities. Two different techniques have been used to assess the main determinants of domestic water in the Sedrata City, North-East Algeria: principal component analysis (PCA) and artificial neural networks (ANNs). To create the ANNs models based on the PCA, twelve explanatory variables are initially investigated, of which nine are socio-economic parameters and three physical characteristics of building units. Two optimum ANNs models have been selected where correlation coefficients equal to 0.99 in training, testing and validation phases. In addition, results demonstrate that the combination of socio-economic parameters with physical characteristics of building units enhances the assessment of household water consumption.