Using Affinity Set and Data Mining to Judge the Necessity of Head Computer Tomography for Patients

• Main Authors: Wei-Chen Ho, 何韋蓁
• Other Authors: Kai-i Huang
• Format: Others
• Language: zh-TW
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/88543919614490014073

碩士 === 大葉大學 === 工業工程與科技管理學系 === 96 === According to statistics from the Department of Health, cerebral vascular disease had been the second-largest cause of death in Taiwan over the last few years, and head injuries leading to death is significant in fatal accidents. In order to diagnose a patient’s brain trauma damage correctly, a head-computed tomography scan (head CT scan) is necessary. In the emergency room, doctors have to determine quickly and correctly whether a patient needs a head CT scan to save his or her life. Therefore, this research employs patients from the Emergency Trauma Department at Chung-Ho Memorial Hospital at Kaohsiung Medical University who have had a head CT scan before as research subjects. This research adopts Affinity Set Theory to construct a model for data mining to deduce important attributes and rules, and to support doctors make judgments. The judgment accuracy rate of the affinity set model proposed in this research is high as 99.4%, higher than those by neural network, logistic regression model, rough set model or support vector machine (SVM) model. Therefore, the model based on the basis of the affinity set can provide doctors decision support to assess whether a head-injury patient needs a head CT scan, so that the diagnosis efficiency of head CT scans can be improved.