cv_two_stage: CV for two-stage

In stephenbates19/logratiolasso: Tools For Large Log Ratio Models

Description

Cross-validation for the lambda and k parameters in the two-stage log-ratio lasso.

Usage

cv_two_stage(z, y, lambda_1 = NULL, k_max = 5, n_folds = 10, ...)

Arguments

z	The input matrix. Assumed to be the logarithmic transform of positive raw inputs.
y	The real-valued response.
lambda_1	Optional vector of lambda values.
k_max	The largest number of log-ratios to consider.
n_folds	Optional number of folds. Default is 10.
...	Additional arguments passed to "two_stage"

Value

mse A matrix of MSE values generated by CV.

best_params The best lambda and k parameter values chosen by CV.

lambda_min A best value of lambda.

k_min A best value of k.

two_step_obj The two_step object fit on the full data.

beta_min The parameter value selected by CV.

lambda The grid of lambda values used.

stephenbates19/logratiolasso documentation built on May 18, 2019, 4:52 p.m.