Experimental Scripts Documentation

This README provides an overview of various scripts used in different experiments. Each script is tailored for a specific dataset or experiment.

Scripts Overview

1. Synthetic.py

• Description: This script is used for experiments on the synthetic dataset.
• Purpose: It corresponds to the first experiment conducted on the synthetic data.

2. Griliches.py

• Description: This script is designed for experiments on the real dataset, griliches76.
• Purpose: It facilitates the experimentations and analysis on the griliches76 dataset.

3. Angrist.py

• Description: This script is used for conducting experiments on the real dataset Angrist.
• Purpose: It is tailored for specific experimental procedures on the Angrist dataset.

4. Approximation_Spam.py

• Description: This script is associated with the efficiency comparison experiment of SHAP.
• Purpose: It is used to compare the efficiency of SHAP approximations in different scenarios.

5. DNNClassifier_Synthetic.py

• Description: Corresponds to the experiments in the appendix involving discrete data.
• Purpose: This script is specifically used for experiments mentioned in the appendix of the study, focusing on discrete data analysis.

6. XGBRegressor_Synthetic.py

• Description: This script is utilized for experiments involving non-neural network models as mentioned in the appendix.
• Purpose: It is designed to facilitate experiments with models like XGBoost Regressor on synthetic datasets as part of extended studies in the appendix.

Additional Information

Each script is integral to the research and provides insights into various aspects of the datasets and models used. For more detailed usage and requirements, refer to the individual script documentation or comments within the scripts.