Description Usage Arguments Value
View source: R/crossValidation.R
Cross-validation for classification models
1 2 3 4 5 6 | CrossValidation(data.train, targetValues, models = c("rf", "glmnet",
"svmRadial", "svmLinear", "gbm", "nnet", "glm"), nFolds = 10,
stratified = TRUE, threshold = NULL, nKeep = NULL, SGoF = NULL,
verbose = FALSE, heatmap = FALSE, PCA = FALSE, extraData = NULL,
tuneKFolds = 2, tuneRepeats = 5, tuneLength = 5, folds = NULL,
precomputedScores = NULL)
|
data.train |
Training data to be divided into folds (must be a data frame) |
targetValues |
Target responses |
models |
list of caret::train models to train |
nFolds |
number of folds |
stratified |
TRUE for stratified folds |
threshold |
list of threshold p-values for selecting features to keep |
nKeep |
list of number of features to keep |
SGoF |
alpha for Sequential Goodness of Fit selection of features |
verbose |
TRUE for verbose output |
heatmap |
TRUE to plot a heatmap of selected features for each fold |
PCA |
TRUE to apply PCA to selected features in each fold |
extraData |
any additional data to add to training data after feature selection |
tuneKFolds |
number of folds for tuning models within each fold |
tuneRepeats |
number of repeats for tuning models within each fold |
tuneLength |
number of parameters to test when tuning models |
folds |
optionally pre-specify which samples go in which fold. Should be
NULL to select folds randomly, or a vector of length
|
precomputedScores |
precomputed scores of features for each fold. A sample of 100 will be tested for each fold and an error thrown in the event of discrepancies. |
A list of predictions for the given model combinations, ready to be passed to CrossValRocCurves
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.