Ridge Regression for Gene Expression Modeling
Usage
ridge_regression(X, G, lambda = 1e-04)
Arguments
- X
A numeric matrix of predictors (cells × features).
- G
A numeric response matrix or vector. Must match number of rows in `X` (cells × genes or 1).
- lambda
Non-negative regularization strength for ridge penalty (default: 1e-4).
Value
A list with components:
- coef
Estimated coefficient matrix (features × genes)
- rss
Residual sum of squares per gene
Details
Performs ridge regression using a closed-form solution. Can be applied to model either a single gene (vector response)
or multiple genes (matrix response) using a shared regularization penalty.
Examples
set.seed(1)
X <- matrix(rnorm(200), nrow = 50)
G <- matrix(rnorm(150), nrow = 50)
fit <- ridge_regression(X, G)
#> Error in ridge_regression(X, G): could not find function "ridge_regression"
str(fit)
#> Error in str(fit): object 'fit' not found