頁籤選單縮合
題 名 | 體檢資料庫探勘--脂肪肝預測模式探討=A Research of Data Mining Applied to the Predictive Model of Fatty Liver |
---|---|
作 者 | 姚志成; 李維平; 吳澤民; | 書刊名 | 資訊科學應用期刊 |
卷 期 | 3:1 2007.06[民96.06] |
頁 次 | 頁113-125 |
分類號 | 312.1、312.1 |
關鍵詞 | 脂肪肝預測; 決策樹; 資料探勘; Fatty liver prediction; Decision tree; Data mining; |
語 文 | 中文(Chinese) |
中文摘要 | 脂肪肝病症不容易察覺,需由專業醫師進行腹部超音波檢查始可發現。如能應用資料探勘分類預測技術於體檢資料庫,藉由挖掘分類規則進行預測,即可輔助診斷決策,節省檢查成本,達到早期發現早期治療的目的。 本研究針對年度體檢資料,藉由決策樹分類方法,分階段進行資料探勘,獲得研究結果如下: 1.中、重度脂肪肝患者之預測準確度高達93%。 2.脂肪肝之重要預測因子包括BMI、三酸甘油脂、ALT(GPT)、腎功能尿酸UA、血液紅血球指數。 3.藉由一般物理檢查項目,包括體重、身高、年齡、血壓、脈博便可初步預測脂肪肝,且預測準確度可達76%。 實驗結果顯示決策樹分類方法運用於醫學體檢資料,可有效找出脂肪肝全般性的關鍵性指標,並建立一個容易閱讀、高準確度且有脈絡可循之脂肪肝預測模式。 |
英文摘要 | It is not easy to perceive the fatty liver disease which is normally diagnosed by the doctor. This research tries to excavate the diagnosing rule of fatty liver disease by using the Decision Tree classification technology to analyze the physical examination data. These diagnosing rules can be used to help fatty liver disease be easily diagnosed by users. Moreover, it will economize the cost of physical examination so that disease and cure can be found in the early days. The results of this research of using the Decision Tree classification technology in analyzing the physical examination data are as follows: 1. Prediction accurate rate is up to 93% for the medium and heavy degree sufferers of fatty liver. 2. The important estimate factors of fatty liver include BMI, triglyceride, ALT (GPT), uric acid UA, red corpuscle index. 3. The factors of weight, height, age, blood pressure and sphygmus can be used to initially predict a fatty liver with accurate rate up to 76%. The experiment results of this research manifest the Decision Tree classification technology to be an effective method in using physical examination data to find out the decisive index of fatty liver. In addition, it can provide easy reading and high accurate prediction model for fatty liver. |
本系統中英文摘要資訊取自各篇刊載內容。