add example with additional attr
This commit is contained in:
parent
4e44dd83a7
commit
78efa13d41
@ -37,3 +37,26 @@ print ('start training with user customized objective')
|
||||
# training with customized objective, we can also do step by step training
|
||||
# simply look at xgboost.py's implementation of train
|
||||
bst <- xgb.train(param, dtrain, num_round, watchlist, logregobj, evalerror)
|
||||
|
||||
#
|
||||
# there can be cases where you want additional information
|
||||
# being considered besides the property of DMatrix you can get by getinfo
|
||||
# you can set additional information as attributes if DMatrix
|
||||
|
||||
# set label attribute of dtrain to be label, we use label as an example, it can be anything
|
||||
attr(dtrain, 'label') <- getinfo(dtrain, 'label')
|
||||
# this is new customized objective, where you can access things you set
|
||||
# same thing applies to customized evaluation function
|
||||
logregobjattr <- function(preds, dtrain) {
|
||||
# now you can access the attribute in customized function
|
||||
labels <- attr(dtrain, 'label')
|
||||
preds <- 1/(1 + exp(-preds))
|
||||
grad <- preds - labels
|
||||
hess <- preds * (1 - preds)
|
||||
return(list(grad = grad, hess = hess))
|
||||
}
|
||||
|
||||
print ('start training with user customized objective, with additional attributes in DMatrix')
|
||||
# training with customized objective, we can also do step by step training
|
||||
# simply look at xgboost.py's implementation of train
|
||||
bst <- xgb.train(param, dtrain, num_round, watchlist, logregobjattr, evalerror)
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user