SDA model trained by tumor gene expression

This function trains a sparse discriminant analysis model using tumor gene expression data with cross-validation for parameter selection. Users can refer to the function sda_model if parameters are known.

Usage

sda_model_cv(
  object,
  stop_vector = NULL,
  lambda_vector = NULL,
  parallel_cores = 1
)

Arguments

object

A CASCAM object.

stop_vector

A vector of negative values with its absolute value representing the number of genes selected for model training.

lambda_vector

A vector of values for the L2-norm for elastic net regression.

parallel_cores

Number of cores for the parallel computation.

Value

A CASCAM object with sda_model, camod_norm_data, tumor_norm_data, tumor_sda_project, and camod_sda_project slots.

References

Clemmensen, Line, Trevor Hastie, Daniela Witten, and Bjarne Ersbøll. 2011. “Sparse Discriminant Analysis.” Technometrics: A Journal of Statistics for the Physical, Chemical, and Engineering Sciences 53 (4): 406–13. https://doi.org/10.1198/tech.2011.08118.

Examples

if (FALSE) {
gene_info <- create_InformativeGenes(tumor_ct, tumor_label2, "ILC")
CASCAM_eg <- create_CASCAM(tumor_aligned, tumor_label, camod_aligned, gene_info)
CASCAM_eg <- sda_model_cv(CASCAM_eg)
}