ibreakdown 0.0.1a4

ibreakdown公司

ibreakdown是具有交互支持的模型无关预测解释器， 库可以显示每个特征在您的预测值中的贡献。在

SHAP或LIME当 ibreakdown还评估本地特性交互。在

算法

算法基于论文“iBreakDown:模型的不确定性”中描述的思想 非加性预测模型“https://arxiv.org/abs/1903.11420和 R（iBreakDown）中的引用实现

算法背后的直觉如下：

The algorithm works in a similar spirit as SHAP or Break Down but is not
restricted to additive effects. The intuition is the following:

1. Calculate a single-step additive contribution for each feature.
2. Calculate a single-step contribution for every pair of features. Subtract additive contribution to assess the interaction specific contribution.
3. Order interaction effects and additive effects in a list that is used to determine sequential contributions.

This simple intuition may be generalized into higher order interactions.

深入的解释可以在算法作者免费书籍中找到： 预测模型：探索、解释和调试https://pbiecek.github.io/PM_VEE/iBreakDown.html

# model = RandomForestClassifier(...)explainer=ClassificationExplainer(model)classes=['Deceased','Survived']explainer.fit(X_train,columns,classes)exp=explainer.explain(observation)exp.print()

特点

• 支持分类和回归的预测解释
• 易于使用的API。在
• 与pandas和numpy
• 支持功能之间的交互

$ pip install ibreakdown

变更