可解释神经网络-the Shapley Additive Explanations (SHAP) method

balabalahoo

已于 2022-09-18 12:22:10 修改

阅读量2k

点赞数

分类专栏：工具文章标签：神经网络人工智能深度学习

于 2022-07-17 17:37:50 首次发布

本文链接：https://blog.csdn.net/m0_46268055/article/details/125834441

版权

工具专栏收录该内容

42 篇文章 ¥9.90 ¥99.00

订阅专栏

超级会员免费看

SHAP（SHapley Additive exPlanations）是Lundberg和Lee于2017年提出的一种解释单个预测的方法，基于博弈论中的Shapley值。该方法包括KernelSHAP和TreeSHAP，提供了对模型解释的新估计方法和全局解释工具。SHAP不仅用于神经网络，还被应用于水文学等领域，通过敏感性分析和特征重要性评估来揭示模型内部工作原理和物理一致性。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

the Shapley Additive Explanations (SHAP) method

SHAP (SHapley Additive exPlanations) by Lundberg and Lee (2017) is a method to explain individual predictions. SHAP is based on the game theoretically optimal Shapley values.

There are two reasons why SHAP got its own chapter and is not a subchapter of Shapley values. First, the SHAP authors proposed KernelSHAP, an alternative, kernel-based estimation approach for Shapley values inspired by local surrogate models. And they proposed TreeSHAP, an efficient estimation approach for tree-based models. Second, SHAP comes with many global interpretation methods based on aggregations of Shapley values. This chapter explains bo

了解本专栏