“…Some of the available methods for formulating data-based approximations and projections include (i) the proper orthogonal decomposition (POD) (a.k.a. principle component analysis (PCA) or the Karhunen-Loève expansion) [26,27,28,29,30,31,32], (ii) the dynamic mode decomposition (DMD) [33,34,35,36,37,38], (iii) deep neural networks (DNNs) [39], (iv) the proper generalized decomposition (PGD) [40], (v) balanced truncation [41] and (vi) reduced basis methods [42].…”