Building Pre-Warning Model For The Oral Cancer Patients’ Medical Prognosis By Discriminant Analysis and Logistic Regression

• Main Authors: Hung-Jui Lin, 林宏儒
• Other Authors: Kun-Ming Yu
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/74430912894876996585

碩士 === 中華大學 === 資訊管理學系(所) === 98 === In 2007, according to the statistical data of Department of Health, Executive Yuan, the Cancer had been won first place from top 10 leading causes of death since twenty-six years ago. According to The World Health Organization (WHO) and Department of Health, Executive Yuan, had report that oral cancer was paid close attention in all kind of cancer. In the past, the oral cancer was incurable terminal illness, but now the oral cancer could be early detected and early treated with advances medical technology. The oral cancer always had a relapse in three years. And if the cancer cells didn’t transfer, the probability of the recovery was very high. In this study, we wanted to solve the problem that detected cancer metastasis as soon as possible, so we employed T test to find a set of medical examinations from all medical examinations that oral cancer patients had been done. It not only can avoid wasted a lot of time on waiting result, but can reduce the cost. And further, we employed Discriminant Analysis and Logistic Regression analysis to develop pre-warning model for the oral cancer patients’ medical prognosis. Finally, we compared the advantages and disadvantages of two models, and then we would select the better one, which is more suitable for assist doctor to make policy. The data of this study was domestic oral cancer patients to make medical examinations in an empirical hospital. The results showed accuracy rate of two pre-warning models were closed 90%. Finally, we compared the effectiveness of the two models and found that each model had its advantages and disadvantages. The results of this study could assist doctors to determine of oral cancer patients prognosis status. It not only can reduce checking time to cause that illness had deteriorated, but also can reduce the wasted of medical resources.