Iterative feature selection method using SHAP values
Project description
IterSHAP: Iterative feature selection using SHAP values
Author: Frank van Mourik, University of Twente
Installation
Install via pip using: pip install itershap
(requires Python version >=3.10,<3.11).
Usage
from itershap import IterSHAP
X, y = get_data() # Replace with data location
fs = IterSHAP() # Create a IterSHAP feature selection object
fs.fit(X, y) # Execute IterSHAP on input data
X_transformed = fs.transform(X) # Only keep the via IterSHAP selected features
Benefits
- Performs well on small high-dimensional datasets
- Guarantees to return a feature subset
- Model-agnostic (limited by shap supported models)
- Validated on synthesised data
- Benchmarked on DEAP dataset
License
Available under the MIT license, which can be found here
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
itershap-0.1.1.tar.gz
(4.8 kB
view hashes)