Methods and data for: On the effect of the progress variable and hyperparameter optimization on ML-FGM accuracy and computational cost

This dataset contains the scripts, data, and methods used to produce the results documented in Chapter 4 of the doctoral dissertation of E.C.Bunschoten titled <em>Consistent Data-Driven Equations of State for Reacting and Nonideal Compressible Fluids</em> . The first set of scripts are used to study the effect of the progress variable definition on the accuracy of deep,dense multilayer perceptrons trained on flamelet data, while the focus of the second part is deriving relations between the hyperparameters, complexity, and accuracy of artificial neural networks used for flamelet-generated manifold applications. The methods are documented in the Python scripts and the data produced following those methods is stored in the compressed zip files.