An Accountable Care Organization (ACO) has engaged us to assist in helping them develop a predictive modeling solution to reduce the prevalence of strokes in the population they manage. Their plan is to develop an outreach program where a team of health coaches will work with members to reduce their risk of stroke. However, they do not have enough resources to engage every member of their population. Therefore, they would like us to develop a model that can be used to match the health coaches with those individuals who have the highest risk of stroke. To solve this problem, we would like you to first build a generalized linear model (GLM). If time permits, we would then like you to try to build a more advanced model to compete with the benchmark performance of your GLM. We are interested to see both the difference in performance and your thoughts on any important advantages or disadvantages to either method.