Interpretable Machine Learning for Insurance



Larry Baeder
Peggy Brinkmann, FCAS, MAAA, CSPA


Machine learning algorithms fit models based on patterns identified in data and can be very complex. In this report, we describe and illustrate a range of methods for interpreting machine learning models from the growing field of Interpretable Machine Learning (IML). We’ve included a file which represents the code used for the research report.


Interpretable Machine Learning for Insurance

Code for Interpretable Machine Learning for Insurance


The authors’ deepest gratitude goes to those without whose efforts this project could not have come to fruition: the volunteers who generously shared their wisdom, insights, advice, guidance, and arm’s-length review of this study prior to publication. Any opinions expressed may not reflect their opinions nor those of their employers. Any errors belong to the authors alone.

Talex Diede, ASA, MAAA
David Evans, FCAS, MAAA
Matthias Kullowatz, ASA, MAAA
Nicholas Hanoian
Kshitij Srivastava
Gabriele Usan
David Wang, FSA, FIA, MAAA

The authors also thank the Project Oversight Group for their diligent work overseeing project development and reviewing and editing this report for accuracy and relevance.

Project Oversight Group members:

Tom Callahan, FSA, MAAA
Michael Destito, FSA, CERA
Annie Girard, FSA, FCIA
Yuanjin Liu, ASA
Vikas Sharan, FSA, FIA, MAAA
Yi Yue Zhang, ASA, ACIA

At the Society of Actuaries:

Korrel Crawford, Senior Research Administrator
Dale Hall, FSA, MAAA, CERA, Managing Director of Research
Mervyn Kopinsky, FSA, EA, MAAA, Senior Experience Studies Actuary
David Schraub, FSA, MAAA, CERA, Senior Practice Research Actuary

Questions or Comments?

Give us your feedback! Take a short survey on this report. Take Survey

If you have comments or questions, please send an email to