10

I am using Gradient boosting for classification. Though the result is improving but I am getting NaN in validdeviance. 

Model = gbm.fit(
  x= x_Train ,
  y = y_Train ,
  distribution = "bernoulli",
  n.trees = GBM_NTREES ,
  shrinkage = GBM_SHRINKAGE ,
  interaction.depth = GBM_DEPTH ,
  n.minobsinnode = GBM_MINOBS ,
  verbose = TRUE
  )

Result

enter image description here

How to tune the parameter to get the validdeviance.

Amarjeet
  • 857
  • 1
  • 9
  • 14

1 Answers1

12

I had the same issue, strangely, we're few on this one ...

Adding train.fraction = 0.5 to the option list solves the issue (it seems there is no default value, and validdeviance is not computed without the train.fraction value explicitly mentioned).

citraL
  • 1,116
  • 10
  • 23