Insurance pricing

Challenging classical econometric models with machine learning


- A few classes of Machine Learning models are (very!) good a discovering non linearities and interesting patterns in data.


- Our engineers used their expertise to discover the optimal factors to estimate the PD and LGD quantities and to apply suitable ML algorithms.


- The results outperformed the existing algorithm and the integration was smooth with the rest of the client's team.

about this project

Insurance companies need as much accuracy as possible when it comes to estimating the probability of default and the loss when the default event happens. This is a central element in fixing the premium (the price of the contract)

Share by: