SciML / DataDrivenDiffEq.jl

Data driven modeling and automated discovery of dynamical systems for the SciML Scientific Machine Learning organization
https://docs.sciml.ai/DataDrivenDiffEq/stable/
MIT License
405 stars 57 forks source link

[WIP] fbDMD (forward and backward DMD function) #314

Closed yewalenikhil65 closed 2 years ago

yewalenikhil65 commented 2 years ago

This PR implements forward and backward DMD based on the papers paper 1 & the second appendix from the paper 2

To-do list includes:

paper 1 may be used for the purpose of tests

AlCap23 commented 2 years ago

Thanks again! I have left some comments and thoughts on your current implementation, which are meant to be helpful. I would especially like to discuss the breakdown of the algorithm as suggested in comment two and hear your opinion on this.

For the test integration we can - as a first step - add the FBDMD() struct in the list of algorithms for all other Koopman tests. If you like, additional tests can be implemented as a new file in ./test/koopman/MeaningfulNameInsertedHere.jl as well ( similar to the current examples ).

yewalenikhil65 commented 2 years ago

Thanks again! I have left some comments and thoughts on your current implementation, which are meant to be helpful. I would especially like to discuss the breakdown of the algorithm as suggested in comment two and hear your opinion on this.

Can you please point me to the comment @AlCap23 ? I cannot see any comments 🤔

AlCap23 commented 2 years ago

I've merged this via #337 and gave credit in the PR.