Bayesian Underwriting of Sinkhole Exposure

Sinkholes have emerged as a significant cost for homeowners and dwelling insurance in Florida. The conditions contributing to the formation of a sinkhole are generally similar over a local area. Thus the observation of known sinkholes and their proximity to a particular area provides a Bayesian predictor of the probability of a future insured claim. Measuring the relative increase in the number of sinkholes and the closeness of these sinkholes to the potential risk gives an indicator of the increased risk.

Keywords: Bayesian Methods