Model module to fit a very large number of machine learning models.
MachineLearn(.df, method = "glm", tuneLength = 8, metric = "ROC",
number = 5, repeats = 1, tuneGrid = NULL, trControl = NULL, ...)
Arguments
- .df
- Internal parameter, do not use in the workflow function.
.df is data frame that combines the occurrence
- method
- The machine learning method to use. Common examples are "glm",
"nnet", "gbm", "glmnet", "rf". See
http://topepo.github.io/caret/modelList.html for a full list. Only
classification or dual use models are useable.
- tuneLength
- How many values of each tuning/hyperparameter should be tried?
- metric
- a string that specifies what summary metric will be used to
select the optimal model. Options are "ROC", "Accuracy" and
"Kappa".
- number
- How many folds to use in cross validation.
- repeats
- How many times should the entire cross validation process be
repeated. Increasing this will reduce instability in your model performace,
but will take longer to run.
- tuneGrid
- Explicitely pass a data frame of tuning/hyperparameter combinations. If
NULL, tuneLength will be used instead.
- trControl
- A named list of further arguments to pass to trainControl. See
trainControl for details.
- ...
- Other arguments passed to
train.
Date submitted
2015-11-13
Data type
presence/absence
See also
caret::train trainControl
Other model: BiomodModel, GBM,
LogisticRegression, MaxEnt,
MaxLike, MyMaxLike,
OptGRaF, QuickGRaF,
RandomForest,
StochasticLogisticRegression,
mgcv