Replies: 1 comment
-
I think it can be a great addition, particularly in terms of generalization. We should pay attention not to breaking anything and add more tests by passing a different Y to all the DMD versions. Probably a possible solution to the issue you raised could be to save an internal flag so that other methods know whether Y was provided or not. |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hey @mtezzele @ndem0 @fandreuz 😃
I propose that we update the$Y$ in the DMD problem $Y = AX$ . Something like this could be extremely helpful, especially since it's common to want to use DMD with time derivatives of the state on the lefthand side, or even a set of snapshot pairs that aren't necessarily from the same time-series.
fit
method for most modules to include an optional argument for a second data matrixI imagine the code could begin with something along the lines of...
The only thing that concerns me about such an implementation is the fact that future usage of
self.snapshots
, such as in_translate_eigs_exponent
,_optimal_dmd_matrices
, or _compute_amplitudes, would potentially yield unexpected behavior.If this would be too invasive of an implementation, we need not bother with it too much. But if anyone has any thoughts on this or also thinks this is a good idea, I'd appreciate the support and I'd love to make this happen. :)
Beta Was this translation helpful? Give feedback.
All reactions