When stakes are high: Balancing accuracy and transparency with Model-Agnostic Interpretable Data-driven suRRogates
作者：
Highlights：
• Procedure to develop an interpretable global surrogate for a complex system.
• Surrogate closely approximates a black box model regarding accuracy and fidelity.
• Automatic feature selection, segmentation and both global and local explanations.
• Satisfy transparency needs of a strictly regulated industry or high-stakes decision.
• Case study on insurance claim frequency prediction for six public datasets.
摘要
•Procedure to develop an interpretable global surrogate for a complex system.•Surrogate closely approximates a black box model regarding accuracy and fidelity.•Automatic feature selection, segmentation and both global and local explanations.•Satisfy transparency needs of a strictly regulated industry or high-stakes decision.•Case study on insurance claim frequency prediction for six public datasets.
论文关键词：Feature selection,GLM,Global surrogate,Insurance,Segmentation,XAI
论文评审过程：Received 10 December 2020, Revised 23 October 2021, Accepted 9 April 2022, Available online 20 April 2022, Version of Record 10 May 2022.
论文官网地址：https://doi.org/10.1016/j.eswa.2022.117230
