top of page



Craig Franze, George Hauser, Haidong Tian, Mita Banik


In this challenge we were presented with a set of simulated pharmacy claims data-billing claims that were run from a pharmacy to a third-party payer who covers some portion of the prescription drug price on behalf of a patient. We used a random forest model to predict the copayments required of patients ahead of time using this claim billing data. We also developed a method to distinguish three types of insurance plans and determine the formulary status, or tier, of each drug with respect to those plans using SVM and KMeans clustering. The drugs generally fall into four tiers under a given plan as a result.

Screen Shot 2022-06-03 at 11.31.35 AM.png
github URL
bottom of page