Abstract
Introduction: Patient no-shows are defined as patients who missed outpatient appointments, either for diagnostic or clinic tests. Identifying those patients is necessary for clinicians and healthcare settings to utilize the resources and improve healthcare efficiency appropriately. This research paper aims to develop a predictive model based on machine learning algorithms to predict patients' failure to attend scheduled appointments. A public data set was divided into training and testing data sets. Two machine learning algorithms, namely decision trees and AdaBoost, were evaluated based on Precision, Recall, True Positive Rate, False Negative Rate, F-measure, and Receiver Operating Characteristic (ROC). Results showed that the decision tree outperformed AdaBoost. The most significant predictors were age and lead time.