Case study card:
Actuarial risk prediction
In insurance, getting pricing right is a balance between two agendas: the first is to ensure pricing is competitive and attracts consumer interest; the second is to ensure the policy is profitable across the entire portfolio of consumers who sign up for the policy.
New policies are by definition predictive guesses about the level of risk and return each segment represents and percentage point changes in predictive accuracy translate to huge profits or losses.
Our academic approach to algorithm selection lead to an improvement in pricing policy optimization and thus improved the core risk-adjusted margin.