查詢結果分析
相關文獻
- 支持向量迴歸技術應用於時變邊界熱通量逆向估算
- 基於PCA的支持向量迴歸分析於房價預測
- 基於支持向量迴歸及支持向量資料描述之半導體CVD製程虛擬量測研究
- Genetic Programming for Classification of Remote Sensing Data
- 逆向熱傳導問題之分析與實驗
- 空間配置規則的推導--人工智慧中案例學習法之運用
- Comparison of ID3 Symbolic Learning and Neural Network for Selecting Deep Foundation Type
- 地理資訊擷取之最佳化方法--以道路養護為例
- 資料採礦中決策樹的理論及應用
- 運用支持向量機與類神經網路於金融時間序列之預測與比較
頁籤選單縮合
題名 | 支持向量迴歸技術應用於時變邊界熱通量逆向估算=Application of Support Vector Regression to Estimate Time Dependent Heat Flux Boundary Condition |
---|---|
作者姓名(中文) | 鄧世剛; 林聖義; 黃宇川; | 書刊名 | 中正嶺學報 |
卷期 | 41:1(A) 2012.05[民101.05] |
頁次 | 頁119-128 |
分類號 | 440.12 |
關鍵詞 | 支持向量迴歸; 機器學習; 逆向熱傳導問題; Support vector regression; Machine learning; Inverse heat conduction problem; |
語文 | 中文(Chinese) |
中文摘要 | 本研究提出以支持向量迴歸演算法求解逆向熱傳導問題的新概念,應用於估算未知時變邊界條件,藉由機器學習演算法優勢解決複雜的逆向估算問題。研究先用模擬分析方式找出逆向關係中的重要變數,接著以支持向量迴歸演算法訓練獲得具備逆向估算能力的迴歸模型。有別於傳統求解方式需詳細逆向推導及繁瑣遞迴迭代才能逼近求解,此新概念僅找出逆向關係中的重要變數,其它複雜計算則交由支持向量迴歸演算法完成,最終的逆向迴歸模型是學習網路透過適當訓練後得。研究結果顯示以支持向量迴歸演算法獲得的逆向迴歸模型,除具備運用便利的優勢外,其估算結果良好、穩健並可克服不適定問題,且有良好泛化能力。 |
英文摘要 | This research provides a new ideal that applying Support Vector Regression (SVR) algorithm to resolve the Inverse Heat Conduction Problems (IHCP) – estimating time dependent heat flux boundary condition. The purpose of this study is to utilize the advantages of machine learning method to deal with the complex inverse problem. We first analysis the direct problem to get the important variables of the inverse relationship and then use SVR algorithm to obtain an optimal regression model, which represents the inverse relationship between the heat flux and the corresponding temperature data. Unlike other transitional methods for IHCP, this new method can get optimal inverse regression model by suitable training process and does not need complex calculation and iterations. Four unknown time dependent heat flux conditions were estimated and the results show that the regression model not only has good convergence and comprehensive ability but also can overcome the ill-posed problem. |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。