查詢結果分析
來源資料
相關文獻
- 雲端資料探勘決策分析平臺
- 類神經網路與資料探勘技術在醫療診斷之應用研究
- Research and Application of Machine Learning: A Case of Iris
- Genetic Programming for Classification of Remote Sensing Data
- 以知識探索為本之知識組織方法論及研究分析
- 整合鑑別分析與類神經網路在資料探勘上之應用
- A Method of Fuzzy-Neural Network to Pattern Classification
- 應用資料探勘技術發掘最適性之線上拍賣競標者
- 乳房X光影像中微鈣化之強化、特徵萃取及辨識
- 資料探勘的技術與應用
頁籤選單縮合
| 題 名 | Research and Application of Machine Learning: A Case of Iris=機器學習之研究與應用:以Iris為例 |
|---|---|
| 作 者 | 林建谷; 陳祐祥; 林玉華; | 書刊名 | 修平學報 |
| 卷 期 | 49 2025.03[民114.03] |
| 頁 次 | 頁73-86 |
| 分類號 | 312.831 |
| 關鍵詞 | 資料探勘; 機器學習; 分類; Data mining; Machine learning; Classification; |
| 語 文 | 英文(English) |
| 中文摘要 | 資料探勘是值得深入研究的領域,機器學習隨著演算法與軟體技術的不斷發展而興盛。人工智慧的機器學習領域的應用也變得更寬廣,預測準確性將隨著存取資料量的增加而提高。本研究以全球公共資料庫中的Iris資料集為例,並使用資料探勘的三個分類器,比較百分比分割和K折交叉驗證的各種參數。最後分析三種分類器的準確性,將準確性最高的分類器用於辨識不同品種的鳶尾花。本研究結果:(1)Random Forest 50/50(97.2%)、Random Forest 60/40(96.6%)和Decision table 50/50(95.9%)是百分比分割的準確率排名前三名。Random Forest 10-fold(94.6%)、Random Forest 15-fold(94.6%)和Random Forest 5-fold(93.9%)是K折交叉驗證最佳準確率的前三名。(2)透過兩種驗證方法的結果,Random Forest是最有效的鳶尾花分類器。研究貢獻可提供給有興趣研究機器學習和分類領域的研究人員作為參考。 |
| 英文摘要 | Data mining is a constantly changing field, and software for machine learning has developed along with its technology. Artificial intelligence includes machine learning, and its uses will only become better. Accuracy will rise in tandem with the amount of data accessed. The Iris data set from the global public database is used as an example in this study. Three classifiers are chosen using data mining software, and various parameters of percentage split and K-fold cross-validation are compared. The accuracy of the classification results is determined. A good classifier for identifying different species of Iris. Findings from the research: (1) Random Forest 50/50 (97.2%), Random Forest 60/40 (96.6%), and Decision table 50/50 (95.9%) have the top three percentage split accuracy percentages. Random Forest 10-fold (94.6%), Random Forest 15-fold (94.6%), and Random Forest 5-fold (93.9%) have the top three accuracy rates in K-fold cross-validation. (2) Random Forest is the most effective Iris classifier when the outcomes of the two verification techniques are combined. Researchers working in the fields of machine learning and classification might use the study findings as references. |
本系統中英文摘要資訊取自各篇刊載內容。