Comparison of data-driven methods in the prediction of hydro-socioeconomic parameters

One of the weaknesses of water resources management is the neglect of the nonstructural aspects that involve the most important relationships between water resources and socioeconomic parameters. Particularly, socioeconomic evaluation for different regions is crucial before implementing water resources management policies. To address this issue, 14 countries in the world that have continuous increasing trends of using renewable water per capita (RWPC) during 1998-2017 were used for the estimation of eight socioeconomic parameters associated with four key indicators (i.e., economy, demographics, technology communication, and health sanitation) by using four different data-driven methods, including artificial neural networks, support vector machines (SVMs), gene expression programming (GEP), and wavelet-gene expression programming (WGEP). The performances of the models were evaluated by using correlation coefficient (R), root-mean-square error (RMSE), and mean absolute error (MAE). It was found that the WGEP model had the best performance in estimating all parameters. The mathematical expressions for these socioeconomic parameters were explored and their potential to be expanded in different spatial and temporal dimensions was assessed. The derived equations provide a quantitative means for the future estimation of the socioeconomic parameters in the studied countries.