https://github.com/bin-cao/shapev
[JMI 2022 | Aggregate 2025] A package for searching equivalent values based on joined SHAP values
Science Score: 49.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
○CITATION.cff file
-
✓codemeta.json file
Found codemeta.json file -
✓.zenodo.json file
Found .zenodo.json file -
✓DOI references
Found 1 DOI reference(s) in README -
✓Academic publication links
Links to: wiley.com -
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (10.7%) to scientific vocabulary
Repository
[JMI 2022 | Aggregate 2025] A package for searching equivalent values based on joined SHAP values
Basic Info
Statistics
- Stars: 3
- Watchers: 1
- Forks: 0
- Open Issues: 0
- Releases: 0
Metadata Files
README.md
ShapEV | Paper : JMI 2022 & Aggregate 2025
A package for identifying Equivalent Value based on joint SHAP values.
ShapEV is an python package that identifies Equivalent Value among features based on cooperative game theory. By leveraging SHAP values, ShapEV decomposes feature contributions and their interactions, defining joint SHAP values to capture combined feature effects.
An Equivalent Value is proposed based on the joint SHAP value, reflecting each feature's overall contribution to the regression target. ShapEV applies this equivalent value to represent the collective behavior of interacting features, allowing users to substitute original individual features with a unified Equivalent Value.
On the modified dataset, ShapEV calculates the contribution of the proposed Equivalent Value by comparing it to SHAP contributions. This validation confirms that, within the SHAP framework, the equivalent value linearly correlates with the target when showing high correlation.
Installation
Install ShapEV using pip:
bash
pip install ShapEV
Usage Example
```python from ShapEV import EVsearch
Instantiate the ShapEV model
model = EVsearch.EVkit('T91-LBE-Train.csv')
Train the model and calculate joint SHAP values
modeltype = 'XGBoost' # Options: 'GradientBoosting', 'RandomForest', 'LightGBM', 'XGBoost' model.fit(modeltype)
model.shap(SPACE=7) # SPACE (int): Maximum number of combinations to return. ```
The ShapEV package simplifies the process of identifying and validating equivalent values, allowing for a more efficient representation of feature contributions in regression models.
About
Maintained by Bin Cao. Please feel free to open issues in the Github or contact Bin Cao (bcao686@connect.hkust-gz.edu.cn) in case of any problems/comments/suggestions in using the code.
License and Usage
© All rights reserved.
This software is provided for academic and research purposes only. Commercial use is strictly prohibited. Any violation of these terms will be subject to appropriate actions.
Owner
- Name: 曹斌 | Bin CAO
- Login: Bin-Cao
- Kind: user
- Location: Shanghai
- Company: Shanghai University
- Repositories: 5
- Profile: https://github.com/Bin-Cao
Machine learning | Materials Informatics｜Mechanics
GitHub Events
Total
- Watch event: 2
- Push event: 10
- Create event: 2
Last Year
- Watch event: 2
- Push event: 10
- Create event: 2
Packages
- Total packages: 1
-
Total downloads:
- pypi 15 last-month
- Total dependent packages: 0
- Total dependent repositories: 0
- Total versions: 1
- Total maintainers: 1
pypi.org: shapev
A package for identifying Equivalent Value based on joint SHAP values
- Homepage: https://github.com/Bin-Cao/ShapEV
- Documentation: https://shapev.readthedocs.io/
- License: MIT
-
Latest release: 0.0.1
published over 1 year ago