I am running a Multivariate Adaptive Regression Splines model with the earth package in R. When I run the model as an additive model I have no problem with convergence. However, when I allow the model to estimate interactions between hinge functions it becomes too complex and the model fails to converge within 25 iterations.
My question, is there an option to increase the number of max iterations? The documentation is not clear on how to do this but I find it hard to believe that there would not be an option to increase the number of iterations!
Here is an example of the model I am running, the degree = 2 allows for interactions between hinge functions.
mars2 <- earth(
y ~ .,
data = df_train,
trace = 4,
degree = 2,
glm = list(family=binomial)
)
"earth glm 1: did not converge after 25 iterations"
