Submissions are evaluated on the accuracy score between the predicted and actual observed output.