Abstract
As aging population and unbalanced diet lead to more patients of prostate cancer, we design a commercial insurance to satisfy insurance demand of medical expenses. Our research data provided by National Health Insurance Research Database (NHIRD) have 400,000 people from 1996 to 2012, and find risk factors of prostate cancer from our model. We estimate probability of suffering prostate cancer by binary model: probit model, logit model, and deal with endogenous variable by instrumental variable. We can calculate reasonable premium under each situation by probability and discounted formula.