Author
Sebha University, Libya
Wadi Alsahtti University, Libya
Abstract
In the middle and southern part of Libya, groundwater is the main water source for all uses due to the extremely low rainfall distribution. The excessively rapid population growth in the urban areas, coupled with very high temperatures – especially in summer – have adversely impacted the groundwater quantity. To study the situation, this study aimed to determine key factors affecting the water consumption. Specifically, the study assesses the factors that impact on consumption for urban residential water in the study area, using the relationship equation (liner regression equation) between the variables of the study. The Waikato Environment for Knowledge Analysis (WEKA) software workbench was selected as a predictive modeling alternative to linear regression model. The strategy used to determine the best operational procedure included (ARFF) to define the type of data, discerning the optimal dataset size, iteration setting, predictor subsets, target attributes. The WEKA models and settings were tested using multiple data strategies on all available data and in situ data since 1973 to 2023. The optimal settings and the linear regression models were used to determining the effect of some factors on water consumption in study area, also to compare between the actual and predict data and get the error value and then compare the error value with the Kappa statistic value. The result of the linear regression model showed a relationship and positive effect between the water demand (WD) and (temperature T), (summer temperature ST), (population P) and (urban population UP). At the same time no relationship existed between (WD) and (income IN), (water price WP).