Additional Considerations in Data Science
A series of issue briefs providing a more complete, contextual view of important topics that arise in data science.
The Data Science and Analytics Committee released an issue paper, Defining Big Data. The paper is intended to build a shared language and concepts of big data and associated themes for the actuarial community, regulators, and other interested parties.
This issue brief defines discrimination (including distinguishing between discrimination, unfair discrimination, and unjust discrimination); presents practical methods for testing and monitoring algorithms; provides a regulatory overview of the issue; and identifies considerations for actuaries, algorithm creators, and regulators.
This issue brief examines the key types of data bias that actuaries may encounter and focuses on the kinds of biases found in modeling data and the implications for algorithmic outcomes.
