Is anyone here aware of any research study or industrial application where ML algorithms have been employed to reduce Adverse Selection and Moral Hazard risk in Insurance market?


1 Answer 1


I haven't seen the specifical cases you mention , but I know of machine learning within insurance, and the types of models I've seen being used always depended on the criticality of the data and use of the results.

So a more general answer: if they need to justify all modelling decisions (as is often the case in insurance), they stick to more traditional models, such as logistic regression, tree-based methods and the like. More cutting-edge models are considered black-boxes.

If the project if more about analysing ad-hoc topics and non-core businesses, but which could perhaps support core business decisions (or otherwise), then anything is game: from stacked bidirectional LSTMs on word embeddings of their contracts to capsule networks on images scraped from social media websites.

I suppose you could model Moral Hazard similarly to fraud, in which case there are plenty of examples to be found online. I think there was even a Kaggle competition on that.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.