I'm trying to apply glmnet and then fixedLassoInf() to some random-effects linear model.
I'm looking at the example in the man page of fixedLassoInf() to figure how to pre-process the experimental design matrix.
But this example uses some randomly generated data experimental design matrix, so I'm not sure whether what I'm doing is right (For some input data I get strange results, even though I try to raise the precision of confidence interval calculation).
It would be nice if there would also be an example showing how to apply it to the more realistic datasets: how to scale and center the real experimental design matrix, how to post-process the output of glmnet and fixedLassoInfo() to get the correct estimates and p-values for the original problem.
(Well, it would be even nicer if the functions of selectiveInference would take care of doing all necessary transformations of the original problem behind the scene).
I'm trying to apply glmnet and then fixedLassoInf() to some random-effects linear model. I'm looking at the example in the man page of fixedLassoInf() to figure how to pre-process the experimental design matrix. But this example uses some randomly generated data experimental design matrix, so I'm not sure whether what I'm doing is right (For some input data I get strange results, even though I try to raise the precision of confidence interval calculation). It would be nice if there would also be an example showing how to apply it to the more realistic datasets: how to scale and center the real experimental design matrix, how to post-process the output of glmnet and fixedLassoInfo() to get the correct estimates and p-values for the original problem. (Well, it would be even nicer if the functions of selectiveInference would take care of doing all necessary transformations of the original problem behind the scene).