Closed antono0106 closed 6 months ago
@Jacob-Stevens-Haas sorry for disturbing, could you please help me to see what’s wrong with my code in order to resolve the issue? Thank you.
Hey @antono0106, I'd love to. I'm going to be spending a bit less time on github over the holidays, however, and SINDy-PI is the part of the codebase I'm least familiar with. So I'm going to ask a bit more rather than diving fully in:
1) Can you get SINDy-PI to work with your code, but on a simpler system? E.g. with $\epsilon=0$, or fewer terms?
2) did you mean to plot x1
equation as model X
and x0_dot
as model_derivative_X'
? Likewise, looks like you swapped x0
and x1
in your system and system derivative
3) plt.legend()
and use a color other than yellow so I don't have to scroll between the plot and the code to read the plot.
4) You have models with the correct terms, they just have too many other terms. Increase your optimizer threshold. There's no great way to "jump" to the correct model in SINDy-PI AFAIK.
Hey again @Jacob-Stevens-Haas , thanks for the answer and sorry for bothering you before Christmas.
I think that in case when there is no "jump" ability using this library, such a way is not working for me.
Thank you again! Have a good holidays! May all your days be merry and bright this season.
I'm using PYSINDY library to reconstruct this 2D system:
Which can be rewritten as:
With parameters:
a10=1, a11=2, a12=3, a20=0, a22=-3, b22=0.0125, v=0.992, e1=1, e2=-1
The plot seems to be correct:
But the none of the output models match to the rewritten one:
How can I get a model similar to the original one? Is this library function choosing problem?
Any help would be highly appreciated.
Thank you in advance, Anton.
The code:
Python version: 3.11, PYSINDY version: 1.7.3