Robust Instance-Dependent Cost-Sensitive Classification
_{Simon De Vos, Toon Vanderschueren, Tim Verdonck, and Wouter Verbeke[2023]}

Description

This is the code for the paper on "Robust Instance-Dependent Cost-Sensitive Classification".

Contact the author at simon.devos@kuleuven.be.

Instructions

Data:

The creditcard transaction data can be found here: https://www.kaggle.com/datasets/mlg-ulb/creditcardfraud
The .csv file should be placed in the data folder as "data\Kaggle Creditcard Fraud\creditcard.csv". You can replace the now empty creditcard.csv file.

Run code:

Run overview.py to execute the experiments as described in the paper.
Settings can be adapted in overview.py:

• Set DIR variable to your custom result folder
• Specify experimental configuration
• Default settings:
  5-fold cross-validation, 2 repeats
  Toy example on synthetic data is generated and displayed
  Three classifiers are trained: logit, cslogit, r-cslogit
  evaluators: traditional, AUC, Savings

Acknowledgments

The code for cslogit is a Python version of the original cslogit by Sebastiaan Höppner et al..

Citing

Please cite our paper and/or code as follows:

De Vos, Simon, Toon Vanderschueren, Tim Verdonck, and Wouter Verbeke. 2023. “Robust Instance-Dependent Cost-Sensitive Classification.” Advances in Data Analysis and Classification, January. https://doi.org/10.1007/s11634-022-00533-3.