Hi all, the default grid for the xbgTree model of the current dev version 6.0-75 is specified differently than the active Cran release 6.0-73. In version 73, _min_childweight and gamma and the only hyperparameters that are held constant, while only the _maxdepth hyperparameter is varied from 1-3 in the dev version. Is this a planned operational change or an unintended effect?
Caret xgbTree Test
Running Cran Caret 6.0-73
library(caret)
## Warning: package 'caret' was built under R version 3.3.3
## Loading required package: lattice
## Loading required package: ggplot2
set.seed(1)
dat <- twoClassSim(100)
X <- dat[,1:5]
y <- dat[["Class"]]
model_class <- train(
X, y, method='xgbTree',
metric='ROC',
trControl=trainControl(
method="cv",
number=5,
classProbs=TRUE,
summaryFunction=twoClassSummary)
)
library(caret)
set.seed(1)
dat <- twoClassSim(100)
X <- dat[,1:5]
y <- dat[["Class"]]
model_class <- train(
X, y, method='xgbTree',
metric='ROC',
trControl=trainControl(
method="cv",
number=5,
classProbs=TRUE,
summaryFunction=twoClassSummary)
)
print(model_class)
## eXtreme Gradient Boosting
##
## 100 samples
## 5 predictor
## 2 classes: 'Class1', 'Class2'
##
## No pre-processing
## Resampling: Cross-Validated (5 fold)
## Summary of sample sizes: 80, 80, 81, 80, 79
## Resampling results across tuning parameters:
##
## max_depth ROC Sens Spec
## 1 0.7954474 0.8186813 0.5857143
## 2 0.7889587 0.8186813 0.6142857
## 3 0.7984825 0.7736264 0.6380952
##
## Tuning parameter 'nrounds' was held constant at a value of 50
## 0.6
## Tuning parameter 'min_child_weight' was held constant at a value of
## 1
## Tuning parameter 'subsample' was held constant at a value of 0.5
## ROC was used to select the optimal model using the largest value.
## The final values used for the model were nrounds = 50, max_depth = 3,
## eta = 0.3, gamma = 0, colsample_bytree = 0.6, min_child_weight = 1
## and subsample = 0.5.
Hi all, the default grid for the xbgTree model of the current dev version 6.0-75 is specified differently than the active Cran release 6.0-73. In version 73, _min_childweight and gamma and the only hyperparameters that are held constant, while only the _maxdepth hyperparameter is varied from 1-3 in the dev version. Is this a planned operational change or an unintended effect?
Caret xgbTree Test
Running Cran Caret 6.0-73
Session Info:
Running Dev Caret 6.0-75
Here only the maximum depth is varied from 1-3:
Session Info: