Dynamic Mode Decomposition (DMD) is an emerging dimensionality reduction algorithm. If you've heard of Principal Component Analysis (PCA) you might know that it is primarily used to reduce the dimensionality of the data before it is fed into some kind of machine learning or statistical model. DMD serves the similar purpose but mainly used for spatio-temporal data. DMD is a purely data-driven (i.e. equations-free) technique that works by extracting the dominant spatio-temporal coherent structures (known as modes) from the data which can then be used to reconstruct/represent the original data at a reduced dimension or make a future-state prediction of the system (with limited accuracy).
Code is written in both Matlab and python. Files will be edited soon and more files will be added.