Generate partial dependence plots

Partial dependence plots (PDPs) help you to visualize the relationship between a subset of the features and the response while accounting for the average effect of the other predictors in the model. They are particularly effective with black box models like random forests, gradient boosting, etc.

autoPDP(
  train,
  trainedModel,
  target,
  feature,
  sample = 0.5,
  modelname,
  seed = 1991
)

Arguments

train	[data.frame \| Required] training sample used to train ML model
trainedModel	[model object \| Required] the object holding the machine learning model and the data
target	[character \| Optional] target variable name. Specify target variable if model object is other than MLR or driveML
feature	[character \| Optional] the feature name for which to compute the effects
sample	[numeric \| Optional] percentage of sample to be considered for training set for faster computation. Default of 0.5
modelname	[character \| Optional] specify whcih model to be plotted
seed	[integer \| Optional] random seed number. Default is 121

Value

List object containing a plot for each feature listed.

Examples


# \donttest{
#' ## Example using DriveML model object
mymodel = heart.model
pdp_chol = autoPDP(heart, mymodel, feature = "chol", sample = 0.8, seed = 1234)
#> input model object is from DriveML

# Type one MLR package
mod <- mlr::train(makeLearner("classif.ranger"), iris.task)
cc = autoPDP(iris, mod, feature = c("Sepal.Length","Sepal.Width","Petal.Length",
                                   "Petal.Width"), sample = 1, seed = 121)
#> input model object is from MLR
# Type 2 DrvieML object
hearML <- autoMLmodel(heart,  target = "target_var",  testSplit = 0.2,
tuneIters = 10,  tuneType = "random",
models = "all", varImp = 20,  liftGroup = 50, positive = 1, seed = 1991)
#> [19:10:12] WARNING: amalgamation/../src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
#> [19:10:12] WARNING: amalgamation/../src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
#> [19:10:13] WARNING: amalgamation/../src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
#> [19:10:13] WARNING: amalgamation/../src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
#> [19:10:13] WARNING: amalgamation/../src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
#> [19:10:13] WARNING: amalgamation/../src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
#> [19:10:13] WARNING: amalgamation/../src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
#> [19:10:14] WARNING: amalgamation/../src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
#> [19:10:14] WARNING: amalgamation/../src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
#> [19:10:14] WARNING: amalgamation/../src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
#> [19:10:14] WARNING: amalgamation/../src/learner.cc:1095: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'binary:logistic' was changed from 'error' to 'logloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
cc = autoPDP(heart, hearML, feature = "chol", sample = 0.8, seed = 1234)
#> input model object is from DriveML

cc1 = autoPDP(heart, trainedModel,target = "target_var", feature = "chol",
sample = 1, modelname = "logreg", seed = 121)
#> Error in paste0(class(trainedModel), collapse = "_"): object 'trainedModel' not found

# Type 3 other ML object
library(randomForest)
#> Warning: package 'randomForest' was built under R version 4.0.4
#> randomForest 4.6-14
#> Type rfNews() to see new features/changes/bug fixes.
#> 
#> Attaching package: 'randomForest'
#> The following object is masked from 'package:ggplot2':
#> 
#>     margin
library(MASS)
rf = randomForest(medv ~ ., data = Boston, ntree = 50)
cc = autoPDP(Boston, rf,target = "medv", feature = "nox", sample = 1, seed = 121)
#> input model object is from other ML pacakges
# }

Arguments

Value

See also

Examples