| Summary: | 碩士 === 中原大學 === 工業與系統工程研究所 === 104 === This research reviews the data fitting methods in the literature and summarizes the common probability distributions in the medical industry. Then, this research develops a new data fitting process in order to determine the probability distribution for each kind of monthly patients’ service time. Based on the proposed data fitting process, this study analyzes the two-year computed tomography (CT) patients’ service time, classifies CT patients into six categories, and determines the probability distribution for each kind of the monthly CT patients’ service time. Next, this study uses mean square error (MSE) and the box-plot to determine which monthly CT patients’ service time can represent the annual CT patients’ service time. Based on the numerical results, this research discusses the possible reason of why some of the monthly CT patients’ service time data are close to the annual CT patients’ service time data. In addition, this study interviews with the hospital manager to figure out the reasons that affect the fluctuation of the monthly CT patients’ service time.
The findings of this research show that the proposed data fitting process can not only replace the traditional process, which is based on the fixed probability distribution or the average patients’ service time, but also solve the data-fitting problem failing to find an appropriate probability distribution. The results can help decision makers select the suitable monthly CT patients’ service time for representing the annual CT patients’ service time in a system simulation model. If the monthly CT patients’ service time meets the following conditions, it will be able to represent the annual CT patients’ service time: 1. There are enough samples for that monthly CT patients, and the range of each CT patients’ service time group is within an acceptable range; 2. The minimum of the monthly CT patients’ service time is less than 3.5 minutes; and 3. The range of each CT patients’ service time group is close to a specific value. Finally, this research indicates that the appropriate monthly CT patients’ service time for representing the annual CT patients’ service time cannot be used to justify the quality of the monthly CT patients’ service time.
|
|---|
