Combining GLMs with GBMs for the Best of Your Pricing Models

In the ever-evolving landscape of P&C insurance pricing, actuaries are increasingly challenged to strike a balance between model transparency and predictive accuracy. This presentation explores how two leading methodologies - Generalised Linear Models (GLMs) and Gradient Boosting Machines (GBMs) - can be leveraged together to achieve just that.

We begin with a brief overview of the P&C pricing process, setting the stage for a deeper dive into GLMs & GBMs: their theoretical foundations, strengths, limitations, and practical applications in pricing. The core of the session focuses on hybrid modelling strategies that combine the stability and interpretability of GLMs with the accuracy of GBMs. Attendees will be introduced to several techniques for integrating these models, along with real-world examples and considerations for implementation.

If you're a pricing professional seeking a trade-off between model performance and interpretable results, this session offers a forward-thinking look at how the future of pricing lies in combining the best of both worlds.