Closed kaibenning closed 3 years ago
You can provide as many lambdas as many penalties you provide. There is a check in the generic Term
class to check that if lambda is iterable and penalties parameter is iterable, then whether they are the same or not. If not the same, it raises a valueError. Also, for constraints, internally a strong lam is provided(1e9
). You can't access this parameter from the user end, neither gam.summary() will output this value.
Thank you! : )
Thank you for this awesome library!
If I set constraints e.g. monotonic decreasing to a term s() of the Linear-GAM, is the penalty on the second-order-derivate still applied? Will I have a for the monotonic constraint and a for the second-order-derivative penalty or just one?
The
gam.summary()
method gives me just []. Is this the expected behaviour? And if i want to can I do both?