查詢結果分析
相關文獻
- 資料探索方法在醫學資料庫之評估
- 資料探勘技術於坡地崩塌之驗證與潛勢評估
- Applying Decision Tree and Neural Network to Raise the Performance of Human Training Quality
- Standard WIP Determination and WIP Balance Control with Time Constraints in Semiconductor Wafer Fabrication
- 應用資料採礦技術分析生物科技新產品之市場區隔:以健康食用油為例
- 信用評等、公司違約率與財務危機預測之探討
- 網路算命使用者行為與特徵分析:資料探勘技術之應用
- 應用機器學習演算法於乳癌資料分析之研究
- 手寫數字辨識模式之建立--結合遺傳演算法與類神經網路
- Form Segmentation and Component Classification for Clinic Document Image Analysis
頁籤選單縮合
題 名 | 資料探索方法在醫學資料庫之評估=The Evaluation of Methods of Data Mining for Medical Database |
---|---|
作 者 | 吳國禎; 趙一平; 蘇振隆; | 書刊名 | 中原學報 |
卷 期 | 30:1 2002.03[民91.03] |
頁 次 | 頁51-61 |
分類號 | 419.21 |
關鍵詞 | 資料探索; 貝氏網路; 決策樹; 倒傳遞類神經網路; Data mining; Bayesian network; Decision tree; Back propagation neural network; |
語 文 | 中文(Chinese) |
中文摘要 | 隨著電腦技術的進步,以及數位擷像設備的使用,資料庫已經廣泛的使用在醫學領 域中。本研究之目的在於分析目前常應用於醫學領域的資料探索演算法的特色以及使用 範圍限制做一整體評估與比較,並提出建議使各種演算法都能發揮其本身最大的特色。 本論文所分析的演算法共有三種,其中包含貝氏網路、決策樹的C4.5演算法以及倒 傳遞類神經網路。所採用的測試資料共有三類,第一額為細針穿刺細胞檢查資料用以診 斷乳部腫瘤細胞是否為惡性,第二類為中醫舌診影像以診斷是否患有上消化道疾病,第 三類為糖尿病健康管理記錄資料用以決定病患之治療方式。 初步結果顯示:貝氏網路診斷能力有不錯的表現,在診斷乳部腫瘤有94.6%以及在 診斷上消化道疾病也有85.8%準確性。決策樹中的C4.5學習演算法對乳部腫瘤有94.4%J 的準確率而對上消化道疾病診斷只有63.9%的準確率。至於倒傳遞類神經網路,對本研 究所採用的三種測試資料的診斷能力是三種演算法中最高的;其中對乳部腫瘤的診斷有96.0%的準確率,而上消化道疾病診斷則有91.69毛的準確率O |
英文摘要 | With the development of computer science, the database has been widely used in medical database. Results from data mining of database could enrich the physician's experience. Three algorithms which includes Bayesian Network(BN), C4.5 in Decision Tree(DT) and Back Propagation Neural Network(BPN) were evaluated in this study Three classes of data were used as testing data. The first set which was from Fine Needle Aspiration Cytology was used to check the breast tumor. The second set which was from the Tongue Diagnosis Image was used to check the disease of upper GI. The third set is diabetic data which was used to decide types of treatment. The primary results show that BN method performed a good presentation indiagnosis ability. The accuracy is about 94.6% in diagnosing breast tumor and 85.5% in upper GI disorder by using BN method. The C4.5 learning algorithms in DT is able to interpret the diagnosis knowledge and rules. However, the accuracy is 94.4% to diagnosis breast tumor, and only 63.9% in upper GI disorder. The best performance among these three algorithms is BPN. The accuracy is 96.0% in diagnosing breast tumor and 91.6% in diagnosing upper GI disorder, respectively. |
本系統中英文摘要資訊取自各篇刊載內容。