PLS-DA Classification — ls_plsda • libscanR

Partial Least Squares Discriminant Analysis for class prediction.

Usage

ls_plsda(dataset, grouping, n_components = 5, validation = "CV")

Arguments

dataset: A ls_dataset() object.
grouping: Character. Column name in sample_info with class labels.
n_components: Integer. Number of PLS latent variables. Default 5.
validation: Character. "CV" (10-fold), "LOO", or "none". Default "CV".

Value

An S3 object of class libs_plsda with elements model, predictions (tibble with predicted/observed class), confusion_matrix, accuracy, class_labels, n_components, sample_info.

Examples

ds <- ls_example_data("tissue")
# \donttest{
if (requireNamespace("pls", quietly = TRUE)) {
  plsda <- ls_plsda(ds, "tissue", n_components = 3, validation = "none")
  plsda$accuracy
}
#> [1] 0.48
# }