effector is an explainability package for tabular data. It offers a variety of global and regional effect methods, under a unified API.

---

effector is compatible with Python 3.7+. Install it via pip:

pip install effector

---

• Start using effector with the Quickstart.
• Learn more on how to use effector in the API DOCS.
• Learn more about it through our Guides.
• Check out the Examples to see effector in action.

---

If you find effector useful in your research, please consider citing the following papers:

@misc{gkolemis2024effector,
      title={effector: A Python package for regional explanations}, 
      author={Vasilis Gkolemis and Christos Diou and Eirini Ntoutsi and Theodore Dalamagas and Bernd Bischl and Julia Herbinger and Giuseppe Casalicchio},
      year={2024},
      eprint={2404.02629},
      archivePrefix={arXiv},
      primaryClass={cs.LG}
}

---

• Friedman, Jerome H. "Greedy function approximation: a gradient boosting machine." Annals of statistics (2001): 1189-1232.
• Apley, Daniel W. "Visualizing the effects of predictor variables in black box supervised learning models." arXiv preprint arXiv:1612.08468 (2016).
• Gkolemis, Vasilis, "RHALE: Robust and Heterogeneity-Aware Accumulated Local Effects"
• Gkolemis, Vasilis, "DALE: Decomposing Global Feature Effects Based on Feature Interactions"
• Lundberg, Scott M., and Su-In Lee. "A unified approach to interpreting model predictions." Advances in neural information processing systems. 2017.
• REPID: Regional Effect Plots with implicit Interaction Detection
• Decomposing Global Feature Effects Based on Feature Interactions
• Regionally Additive Models: Explainable-by-design models minimizing feature interactions

---

effector implements the following methods:

Method	Global Effect	Regional Effect	Paper
PDP	PDP	RegionalPDP	PDP, ICE, GAGDET-PD
RHALE	RHALE	RegionalRHALE	RHALE, DALE
SHAP-DP	ShapDP	RegionalShapDP	SHAP, GAGDET-DP
ALE	ALE	RegionalALE	ALE, GAGDET-ALE
d-PDP	DerPDP	RegionalDerPDP	d-PDP, d-ICE