Skip to content

SianC/Fair_Model_Test

Repository files navigation

Fair_Model_Test

This code is associated with the fair model testing work

Data creation

Synthteic data creation can be completed using Data_from_DAG.R

Data cleaning/ required documents

These documents are needed to be downloaded for the later work:

  • Conf_Mat.py
  • Results_df_create.py
  • Split_data.py

Model creation

The following documents contain the model creation and gridsearches:

  • DT_model.py
  • KNN_model.py
  • LR_model.py
  • RF_model.py
  • SVM_model.py

Fairness test creation

The following documents contain the fairness notions:

  • Treatment_equality.py
  • Statistical_parity.py
  • Equalised_Odds.py

Do the work!

Run_through_full requires inputting the folder containing the datasets and allows you to run and save the results of all 5 models. Note - the Random Forest and SVM output times taken as they are computationally expensive when there is a large amount of data. Run_trhough_FT requires inputting the folder containing all the results and allows you to run and save the 3 datasets. Graphs.py creates graphs for the results of the fairness tests

About

This code is associated with the fair model testing work

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published