查詢結果分析
來源資料
頁籤選單縮合
題名 | 應用決策樹演算法以探究高科技員工潛在的糖尿病之危險因子=Applying Decision Tree to Explore Risk Factors in Developing Diabetes among Employees in High-Tech Industry |
---|---|
作者 | 陳銘樹; 王建智; 王麗雁; | 書刊名 | 健康管理學刊 |
卷期 | 6:2 2008.12[民97.12] |
頁次 | 頁135-146 |
分類號 | 415.668 |
關鍵詞 | 資料探勘; 決策樹; 第二型糖尿病; 危險因子; Data mining; Decision tree; Type 2 diabetes; Risk factor; |
語文 | 中文(Chinese) |
中文摘要 | 高科技產業的產品生命週期短,競爭激烈,雖然員工平均年齡不高,但多數員工超時工作,高度的工作壓力以及缺乏運動或三餐外食的生活習慣促使高科技員工健康逐漸亮起紅燈。在台灣,糖尿病位居十大死因第四位,且有每年上升的趨勢,第二型糖尿病屬於不易自行察覺的疾病,更容易被一般年輕的科技業員工所忽視。本研究以高科技產業員工爲研究對象,根據員工連續兩年的健康檢查報告資料,應用決策樹分類技術分析員工健康狀況的變化,以找出屬於該族群的糖尿病危險因子。過去文獻中的結果,無法讓目前健康指標還在正常範圍內的民眾知道自身健康是否開始產生不好的變化,只能讓健康狀況已經異常的民眾得知自己罹病的機率。因此本研究提出新的分析概念,將「因子變化量」納入考量,如此將可讓民眾知道糖尿病危險因子的變化量是否在安全範圍內。分析結果顯示,考慮因子變化量的決策樹模型預估的準確率最高,總正確率高達82.51%。分析高科技產業員工資料後發現,包含1.空腹血糖;2.血壓;3.三酸甘油脂;4.肝功能(GPT);5.白血球等五項檢驗數據爲第二型糖尿病的統計上的重要危險因子。 |
英文摘要 | The products life cycle of High-Tech industry is short, the competition is intense. Although the employee average age is not high, most employees work overtime, are under high pressure and lack of exercise, employees' lifestyle causes a danger signal to appear. Diabetes has been ranked the fourth position in the top ten causes of death in Taiwan, but has been rising every year. Type 2 Diabetes Mellitus is often difficult to perceive by oneself, and is commonly ignored by young employees in the High-Tech industry. In this study, according to two consecutives years of health examination from employees in the High-Tech industry, applying Decision Tree classification technology analyzes the extent to which the health of staff changes, in order to find out the diabetes risk factors of workers in this industry. The resulting literature is unable to give present employees in the normal range populace an indicator for when their own health starts to take a turn for the worse; rather, to only inform the healthy of levels of the morbidity rate. So this study presentation promotes a new concept that considers the "quantities of factor changes", it will let the populace know if the change quantities of diabetes risk factors are in a safe range or not. Analysis results reveal, assuming the "quantities of factor changes" of the decision tree model have the highest accuracy, the correct rate of 82.51%. After data analyses of the High-Tech industry staff, 5 factors were determined to put one at risk for Type 2 Diabetes Mellitus. These factors are: (1) impaired fasting glucose, (2) blood pressure, (3) triglyceride, (4) Glutamate pyruvate transaminase (GPT), and (5) white blood cell count (WBC). |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。