Source codes for an integrated machine learning approach for real-time prediction, diagnostics and optimization of uranium-leaching groundwater system

The combined capabilities of a Joint Convolution and Recurrent Neural Network model, alongside a data-driven predictive model fused with the Expected Gradient algorithm, have revolutionized the accurate prediction and explainable diagnostics of uranium concentration in a uranium-leaching system with multiple wells.Further integration with optimization algorithm allows for real-time management of in-situ uranium leaching.