Using Multivariate Analysis to Explore the Relationship between the Health Examination Items and Colon Abnormalities

• Main Authors: Jhih-Da Chen, 陳志達
• Other Authors: Chun-Yuan Cheng
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/34926734685263989510

碩士 === 朝陽科技大學 === 工業工程與管理系碩士班 === 101 === Colorectal cancer is quite common and has high mortality, the incidence of colorectal cancer in Taiwan has increasing trend in recent years. Hence, the seriousness of colorectal cancer should not be underestimated. Therefore, the prevention and screening of colorectal cancer is a very important issue. Since the government of Taiwan provides free health examination service for elder adult people, it is worthwhile to explore the critical items of health examination which are correlated with the colon abnormalities. Then, the early detection of colon abnormalities through health examination can improve the preventive health care. In this study, multivariate analysis method is used to investigate the correlation between health examination items. The highly correlated health examination items are clustered as groups to increase the independency of the health examination items (or items clusters). Then, the discriminate analysis is applied to explore the critical health examination items (or items clusters) which are significantly correlated with colon abnormalities. The results of this research show that accuracy rate of abnormality prediction is 72.4% when the colonoscopy results are divided into 2 class (normal and abnormal), where the critical health examination items are founded as age, alkaline phosphatase, erythrocyte sedimentation rate, triglycerides, and thyroid-stimulating hormone significance.