The following tables contain a short description for each python file. Common refers to files related to both papers in this repository.The next two tables are paper-specific.
| Module | Description |
|---|---|
| credit_utils.py | Preprocesses the credit dataset. |
| fico_utils.py | Preprocesses the lending dataset. |
| utils.py | Contains auxiliary functions. |
| real.py | Performs one experiment on real data with options about counterfactual explanations / full transparency. |
| Module | Description |
|---|---|
| fair.py | Performs one experiment on real data under a matroid constraint. |
| min_cost.py | Finds minimum cost explanations. |
| max_cover.py | Finds diverse explanations. |
| greedy_deter.py | Finds explanations maximizing utility. |
| greedy_rand.py | Finds explanations and a policy maximizing utility. |
| greedy_fair.py | Finds explanations maximizing utility under a matroid constraint. |
| configuration_counterfactuals.py | Contains instance generation functions. |
| Module | Description |
|---|---|
| bruteforce.py | Finds the optimal policy maximizing utility under full transparency. |
| dp.py | Dynamic programming algorithm for finding a close to optimal policy on additive outcome monotonic instances. |
| iterative.py | Iterative algorithm for approximating the optimal policy. |
| thres.py | Iterative algorithm that searches over all threshold policies and picks the one with maximum utility. |
| configuration_optimal.py | Contains instance generation functions. |