Code and Data for "Null importance: Disentangling relevance for interpretable machine learning."

<p dir="ltr">The code here mirrors that in our GitHub repository (github.com/krisrs1128/null_importance). Each case study is a subdirectory of the <code>case_studies</code> folder. The READMEs describe the commands that are used to replicate the analysis and visualization.

Under case studies, the <code>environment.yml</code> file defines the conda environment "interpretability_case_studies" which describes all the all the necessary packages to replicate the results.</p><p dir="ltr">The real and simulated data used as inputs for these case studies are further subdirectories, e.g., <code>sweep_tabular/data</code>. The experiments are run using Hydra, and the associated configuration files are stored under <code>conf/</code>.