effector 0.0.260

No project description provided

Effector

Documenation | Global Effect | Regional Effect | API | Tutorials

Effector is a python package for global and regional effect analysis.

---

---

Installation

Effector is compatible with Python 3.7+. We recommend to first create a virtual environment with conda:

conda create -n effector python=3.7
conda activate effector

and then install Effector via pip:

pip install effector

Methods and Publications

Methods

Effector implements the following methods:

Method	Global Effect	Regional Effect
PDP	PDP	RegionalPDP
d-PDP	DerivativePDP	RegionalDerivativePDP
ALE	ALE	RegionalALE
RHALE	RHALE	RegionalRHALE
SHAP-DP	SHAPDependence	RegionalSHAP

Publications

The methods above are based on the following publications:

• Global Effect:

  • PDP and d-PDP: Friedman, Jerome H. "Greedy function approximation: a gradient boosting machine." Annals of statistics (2001): 1189-1232.
  • ALE: Apley, Daniel W. "Visualizing the effects of predictor variables in black box supervised learning models." arXiv preprint arXiv:1612.08468 (2016).
  • RHALE: Gkolemis, Vasilis, "RHALE
  • SHAP-DP: Lundberg, Scott M., and Su-In Lee. "A unified approach to interpreting model predictions." Advances in neural information processing systems. 2017.

• Regional Effect:

  • 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