Start time: 12.30 pm
End time: 1.30 pm
A recording of this webinar can be found here:
This webinar will be held over zoom. Click here to read the User Guide for Zoom webinars/meetings.
As the application of machine learning models in non-life insurance pricing gained momentum in recent years, the importance of explainability and interpretability increased significantly.
The talk will cover methods aiming to turn black box pricing models into white box ones including:
- Use of rule-based and surrogate models
- Feature importance
- Partial dependence plots (PDPs)
- Individual conditional expectations (ICE)
Practical examples will be presented using Python libraries that enable users to address both global and local model explainability such as LIME (local model explanations), SHAP (feature importance metrics), ELI5 (human readable model interpretations) and explainer dashboards (model explainability dashboards).
Furthermore, methods to assess fairness of non-life pricing models are also important to avoid unintended discrimination and regeneration of biases prevalent in the data. We will demonstrate how the aequitas and fairml libraries in Python can help understand such biases, compare biases between groups (bias disparity) and visualise bias metrics. We will conclude with a discussion on considerations for defining and setting fairness metrics for pricing models.
Bence Zaupper is a managing consultant at Finalyse Dublin. Bence is a qualified actuary with 20 years experience in life and pensions, risk and capital management. He is a member of the Data Analytics Committee of the Society and holds a post-graduate diploma in Data Analytics since 2021. Bence was the coordinator of the Python Study group for members of the Society in 2023.
Marino San Lorenzo is a consultant at Finalyse Warsaw. He is a qualified actuary and data scientist and has 7 years experience in developing non-life actuarial and machine learning models. Marino completed successful projects for insurance companies using machine learning for non-life pricing, fraud detection and prediction of churn and property prices.
Deep Dive: This is a Deep Dive Event.