I am trying to plot variable importance from a support vector model using the vip package in R. However, the plot is empty.
set.seed(2022)
pet_uv_raw_svm_fit %>%
extract_fit_parsnip() %>%
vip(method = "permute",
target = "class",
metric = "accuracy",
nsim = 20,
pred_wrapper = kernlab::predict,
train = pet_uv_train[, -1],
)
With the same code but only different data, then I get a good variable importance plot.
set.seed(2022)
pet_uv_svm_fit %>%
extract_fit_parsnip() %>%
vip(method = "permute",
target = "class",
metric = "accuracy",
nsim = 20,
pred_wrapper = kernlab::predict,
train = pet_uv_train[, -1],
)
So, my question is what does an empty vip plot mean? and what is the reason causing the result?
Any idea is very much appreciated.