This project is concerned with predicting the copayment that is expected of a patient for a medication using claim billing data that records information on the transaction date, pharmacy name, diagnosis, drug details, insurance plan details, claim status, and patient copay. The goals of this project are 3-fold: predict the expected copay of a patient for a particular medication, predict whether the claim will be accepted or rejected by the insurance plan, and cluster medications according to formulary status and copayment requirements.

Our proposed predictive model has 2 stages. In the first stage, a classifier determines whether the claim will be accepted or rejected. If the claim is rejected, the patient copay is $0. Otherwise, a regression model is then used to compute the payment that is expected of a patient.

For our regression model, we achieved 6.7921 MSE and 0.0192 MAPE on the test data set. The low MAPE suggests strong predictive capability of our model given that the copay