Extract coefficients from penalized regression

1
votes

I am using the sparse group lasso, which is a penalized regression. The package I am using is SGL. I tried to run the examples in my R, and the code is given as below

set.seed(1)
n = 50; p = 100; size.groups = 10
index <- ceiling(1:p / size.groups)
X = matrix(rnorm(n * p), ncol = p, nrow = n)
beta = (-2:2)
y = X[,1:5] %*% beta + 0.1*rnorm(n)
data = list(x = X, y = y)
cvFit = cvSGL(data, index, type = "linear")

I tried to extract the regression coefficient of cvFit, but it turns out to be 

coef(cvFit)
NULL

Can anyone tell me what is wrong? Thanks in advance.

1
rregression
See str(cfFit) and determine the exact location of the coefficients. – Roman Luštrik
@ Roman Luštrik what do you mean by determining the exact location of the coefficients? – Xst67
Not familiar with this function but by typing cvFit you can see what values can be extracted. ie cvFit$fit$beta or cvFit$lambdas – user20650
cvFit$fit$beta contains a matrix of coefficients, one column for each lambda in the regularization path. – davechilders
Post output of str function and I'll show you how to access the output. – Roman Luštrik

1 Answers

0
votes

This extracts the coefficients from the model with the minimum lambda value.

coef(fit,s=cvfit$lambda.min)