We start with the business problem:
A model can be accurate, but if it doesn’t provide the information a client requires to make decisions then it’s no good to anyone. We know the importance of speaking to people in terms they are familiar with and use in their day-to-day work. At Claim Analytics, we speak the language of insurance which enables us to ensure the model will fit and support the business process at the outset. And, most importantly, we ensure at the outset on making sure that the model output will fit into and support the business process.
We have the experience to choose the most appropriate model and apply it effectively. Often one model is not sufficient and we may use a composite model we refer to as a ‘committee of experts’. Once the model is complete, we test the model on hold-back data, which the model has not seen before, to ensure that the model’s predictions match the actual results.
Before a model is ready for production, we work with our clients to validate the model. At the outset of the project we ask our clients to retain some of the data (typically 10-20%) from the development process. After completing the model and testing it on our hold back data, the output is tested against the data that the client held. Data that we have never seen. This provides objective proof that the model is both robust and accurately predicts the expected outcomes.
Report Presentation and Action
It’s important to present results in a way that is easy to understand and actionable. We employ experienced insurance business professionals to work with our clients to turn our model results into business results.
Please contact us if we can be of any assistance in helping you improve your company results through predictive analytics:
What our clients say...
“The working relationship we have with the Claim Analytics team is exceptional. They have been fully supportive of our business needs and continue to offer assistance and training.”
Example: Rehabilitation Report
Proactive, automated scoring can be applied to rehabilitation to identify claims that would likely have improved outcomes with rehab intervention.