Claim Analytics use predictive modeling techniques to develop more accurate Workers Compensation Pricing/Underwriting rates. Claim Analytics look back at insurers historical data to determine the rating factors and develop predictive algorithms to provide a sophisticated and highly accurate rating mechanism.

The use of external data feeds can provide an added source of information for an underwriting model, credit rating data, previous claim history, criminal record information etc.

Deliverables
Claim Analytics will provide a model that will produce Workers Compensation Pricing/Underwriting rates. The model will have no constraints on how the rating variables are applied to achieve the final result.

Data Requirements
Claim Analytics will review the availability of historical data and decide on which data elements need to be provided and assess the cleanliness of the data.

Validation of the model
Validating the model provides confirmation that the Claim Analytics pricing methodology provides a better fit on a group-by-group basis to the historical data than the current pricing methodology.

When providing data to Claim Analytics, approximately 10% of data is excluded from all groups. This excluded data, called the “validation sample”, should be a representative sample of the total business.

Once the Claim Analytics pricing model is finalized the sum of the squares of the differences, determined on a reasonable segmentation cut of the business, between actual claim costs and priced claim costs will be calculated for the validation sample.