Introduction to Clustering for Territorial Ratemaking


Territorial ratemaking is a tricky process. It doesn’t work well in traditional regression approaches, but is made easier and more transparent through clustering. In this APEX, Consulting Actuaries Kunshan Yin and Joey Sveda will introduce two common clustering methods – K-means clustering and hierarchical clustering – and discuss how these algorithms can be incorporated into regression-based pricing models for better results. They will also discuss territorial ratemaking in the context of the evolving conversation on disparate impact in insurance pricing.