Skip to contents

Standard error estimation and significance testing for coefficients estimated in linear ridge regression. ridgePvals re-implement original method by (Cule et al. BMC Bioinformatics 2011.) found in ridge-package. This function is intended to use with cv.glmnet output.

Usage

ridgePvals(x, y, beta, lambda, standardizex = TRUE, svdX = NULL)

Arguments

x

input matrix, same as used in cv.glmnet.

y

response variable, same as used in cv.glmnet.

beta

matrix of coefficients, estimated using cv.glmnet.

lambda

lambda value for which beta was estimated.

standardizex

logical flag for x variable standardization, should be set to same value as standarize flag in cv.glmnet.

svdX

optional singular-value decomposition of x matrix. One can be obtained using link[base]{svd}. Passing this argument omits internal call to link[base]{svd}, this is useful when calling ridgePvals repeatedly using same x.

Value

a data.frame with columns

coef

beta's names

se

beta's standard errors

tstat

beta's test statistic

pval

beta's p-values