Feature Selection for Predictive or Prognostic Models Using LASSO Regression

Applies LASSO (Least Absolute Shrinkage and Selection Operator) regression to construct predictive or prognostic models. Supports both binary and survival response variables, utilizing cross-validation for optimal model selection.

Usage

lasso_select(
  x,
  y,
  type = c("binary", "survival"),
  nfold = 10,
  lambda = c("lambda.min", "lambda.1se")
)

Arguments

x

A numeric matrix. Features (e.g., gene symbols or CGI) as row names and samples as column names.

y

A response variable vector. Can be binary (0/1) or survival data (e.g., survival time and event status).

type

Character. Model type: "binary" for binary response or "survival" for survival analysis. Default is "binary".

nfold

Integer. Number of folds for cross-validation. Default is 10.

lambda

Character. Regularization parameter selection: "lambda.min" (minimum mean cross-validated error) or "lambda.1se" (one standard error from minimum). Default is "lambda.min".

Value

Character vector of selected feature names with non-zero coefficients in the optimal LASSO model.

Author

Dongqiang Zeng

Examples

set.seed(123)
gene_expression <- matrix(rnorm(100 * 20), nrow = 100, ncol = 20)
rownames(gene_expression) <- paste0("Gene", 1:100)
colnames(gene_expression) <- paste0("Sample", 1:20)

# Binary response example
binary_outcome <- sample(c(0, 1), 20, replace = TRUE)
lasso_select(
  x = gene_expression,
  y = binary_outcome,
  type = "binary",
  nfold = 5
)
#> Warning: one multinomial or binomial class has fewer than 8  observations; dangerous ground
#> Warning: one multinomial or binomial class has fewer than 8  observations; dangerous ground
#> Warning: one multinomial or binomial class has fewer than 8  observations; dangerous ground
#> Warning: one multinomial or binomial class has fewer than 8  observations; dangerous ground
#> Warning: one multinomial or binomial class has fewer than 8  observations; dangerous ground
#> character(0)