How to add time dependence to candidate function library.

[sdal750-sudo]

I am trying to add a y and yM2 to the polynomial library. y is equal to y0 + epsilont, where epsilon and y0 are fixed and defined. However when i do this, it fits two seperate y's to the model. How do I fix this to have only one y in the candidate function library.

M = np.stack((m1, m2), axis=-1)

poly_library = ps.PolynomialLibrary(degree=3)

y_functions = [
lambda t: y0 + epsilon * t,
lambda x, t: x[1] * (y0 + epsilon * t),

]

y_function_names = [
lambda t: 'y',
lambda x, t: 'y M2',

]

y_library = ps.CustomLibrary(
library_functions=y_functions,
function_names=y_function_names
)

combined_library = y_library

differentiation_method = ps.FiniteDifference(order=2)
optimizer = ps.STLSQ(threshold=0.01)
model = ps.SINDy(
differentiation_method=differentiation_method,
feature_library=combined_library,
optimizer=optimizer,
feature_names=["M1", "M2"],
)

model.fit(M, t=t)
model.print()

[Jacob-Stevens-Haas]

Thanks for your question. I'm not sure I understand. What output of model.print() do you expect? What output do you get?