Artificial intelligence and regression analysis in predicting ground water levels in public administration

Water level prediction of ground water can be considered as a very important tool in water resources management. This research implements artificial neural network models in order to build the optimal forecasting models for predicting the water levels and regression analysis in order to evaluate the prediction accuracy. In this research, an Artificial Neural Network Perceptron is applied in order to construct forecasting models for predicting water levels of ground water. The developed models are then compared each other in order to find the optimal one according the best performance that will lead to the most accurate predictions. Several topologies were tested in order to discover the best forecasting model. The different predictive models were constructed by implementing different number of the nodes in the hidden layers, also by testing different number of the hidden layers. The results showed an increased prediction accuracy of the developed Artificial Neural Network Models. This research is aiming at providing the scientists and engineers with the optimal prediction models in order to be used for forecasting the water levels of ground water with increased accuracy compared to other prediction techniques and methods. The developed forecasting models can provide accurate predictions of water levels of ground water which can be very valuable for the public administrators and stakeholders in water resources management.