Track: Data Analytics

Abstract

Outpatient appointment scheduling problems require predicting outpatients’ treatment or examination time more precisely to reduce patients’ waiting time and physicians’ idle time. As there are many characteristics of outpatients’ examinations, the treatment or examination time of each outpatient is varied and random. Thus, this research applies the data mining method to classify outpatient’s examination time of a case image center. The relevant one-year data of outpatients’ examination time are collected and preprocessed, such as the outpatients’ examination position, the examination time and the number of examination images, the examination performed by radiological technologists, the medical order issued by which department of physicians, etc. This study uses three kinds of classification methods, including decision tree methods (e.g., C4.5, CART, and CHAID), Logistic regression, and artificial neural network to train and verify data. The results show that CART outperforms other methods and extracts five rules to classify outpatients’ examination time.