Client needed an PoC model built on the available data to model probability of a breach for the specific company. Deliverables:
- Predictive models to estimate probability of data breach, possible amount of losses
- Implied risk densities for the companies in given subgroup
- Cramer-Lundberg actuarial model to account for probability of default for the actuarial company given its initial funding and hedge portfolio of clients.
R, h2o, randomforest, actuarial, libraries for estimating implied distributions, xgBoost, regressions.