Code for conducting post-selection inference after lasso selection.
The algorithm is proposed in Liu, S. (2023). An Exact Sampler for Inference after Polyhedral Model Selection.
git clone https://github.com/liusf15/selinf_sampler.git
The proposed SOV sampler is written in cython so it requires compiling:
cd src
python setup.py build_ext --inplace
The vignette provides an example showing how to use the implemented method to construct confidence intervals and compute p-values for variables selected by the lasso. The lasso uses a subset of the data and the inference is performed conditioned on the selection.