Simulating the Interdependence Between Electricity Production and Water Release, Based on a Methodology that Combines Statistical Models and Artificial Intelligence

Abstract

    The research presents a hybrid innovative model that improves the accuracy of the prediction of water release and energy production through the combination of artificial neural networks with econometric modeling techniques. The research assumed a combined approach, taking advantage of the strength of neural networks to discover complex and nonlinear relationships in data and the benefits of the vector regression models in long-term equilibrium relationships. Two integrated hybrid models that address each variable independently while maintaining their interrelationships were created. With notable gains in a number of evaluation criteria on both training and future data, the suggested model outperformed conventional models in terms of prediction accuracy. Utilizing the combination of cutting-edge statistical approaches and artificial intelligence technologies, the created methodology is distinguished by its capacity to handle the intricacies of time series.