Train DeepCC Model — train_DeepCC

This function trains DeepCC Model on the training data using the modern keras3 interface. The network architecture, training recipe, and legacy model format are preserved from DeepCC 0.1.1.

Usage

train_DeepCC_model(
  trainData,
  trainLabels,
  epochs = 100,
  dropout = 0.4,
  activation_func = "selu",
  validation_split = 0.2
)

Arguments

trainData: a data.frame containing functional spectra of training data (each row presents one sample)
trainLabels: a character vector containing lables of training data
epochs: the number of epochs
dropout: dropout rate
activation_func: activation funtion
validation_split: fraction of training data to use for validation

Value

a trained DeepCC model with classifier, levels, and feature_names fields

Examples

if (FALSE) { # \dontrun{
set.seed(42)
eps <- as.data.frame(matrix(rnorm(20*50), nrow=20, ncol=50))
colnames(eps) <- paste0("F", seq_len(50))
labels <- sample(c("A", "B", "C"), 20, replace=TRUE)
deepcc_model <- train_DeepCC_model(eps, labels, epochs=2)
} # }