Model Validation
Be confident that model outputs appropriately capture your risk.
As risks continue to increase in complexity, models have become more sophisticated. Get a better understanding of how well risk models reflect the specific characteristics of your portfolio—at an aggregate or granular level—to ensure that the results you’re using to manage your business are appropriate.
Ensure that your model meets your risk management needs.
Perhaps you target unique risks or a specific geography, or have unique claims handling processes. Detailed claims analysis can provide you with the confidence that your unique portfolio is represented by the model.
Use real-world data to validate your model.
By using your claims data from historical extreme events and tying it to exposure data that correlates to the event timeline—including all the necessary data processing—AIR can perform a thorough model validation using real-world data to better capture how your model reflects your risk.
Validate your model and your exposure data at once.
When conducting a model validation, AIR can use a chosen historical extreme event to do an in-depth evaluation of your exposure data to help you understand any potential gaps or inconsistencies that could impact model performance.
Employ a comprehensive approach to model validation.
Reviewing specific events allows AIR to see how well the modeled damage ratios translate to actual claims experience, and evaluating many years of claims for a particular peril helps determine if the model’s catalog is capturing your risk at an average or regional level.
Optimize your model output.
If the model validation process reveals that modification of the modeled losses would be suitable, AIR can help you structure the modifications appropriately to reflect your unique risks.
