Closed p-schaefer closed 3 years ago
Hello. You would need to specify y = "Resp"
in building the explainer and use a metric compatible with 0/1 predictions (logLoss is a metric used when predictions are probabilities, not classes).
fl <- flashlight(model = m1, data = test1, label = "rf",
metrics = list(AUC = AUC),
predict_function = function(mod, X) predict(mod, X)$predictions,
y = "resp")
plot(light_importance(fl, lower_is_better = FALSE, m_repetitions = 10),
fill = "orange")
I am having trouble running the functions for classifications using Ranger, i.e.,