不同优养水质共线性分析及模式选择之研究Colinearity-水土保持学系.PDFVIP

不同優養水質共線性分析及模式選擇之研究 (1) (2) 陳鴻烈 蔡大偉 摘 要 本研究是以德基水庫優養化趨勢相關重要參數為研究對象，運用不同迴歸分析選擇法篩選 出與優養化相關之最重要因子，以提升迴歸模式的解釋能力，協助水庫優養化管理策略之制 定。研究中共使用三種迴歸分析選擇法，包括順向、逆向選擇迴歸分析與逐步迴歸分析，分別 針對先經過單迴歸分析篩選的優養化重要因子進行分析，以挑選出最適宜的因子。研究結果顯 示，三種分析法的標準化 R2 值均較原始複迴歸表現好，其中又以順向選擇分析最佳，表示使用 迴歸分析選擇法可有效選出與優養化最相關的因子，進而提升模式的解釋能力。 關鍵詞( ：優養化、複迴歸分析、參數選擇、水質因子、德基水庫) Colinearity Analysis and Model Selection of Water Quality Factors Related to Eutrophication Paris Honglay Chen Professor, Department of Soil and Water Conservation, National Chung-Hsing University, Taichung, Taiwan 402, R.O.C. David D-W. Tsai Graduate Student, Department of Soil and Water Conservation, National Chung-Hsing University, Taichung, Taiwan 402, R.O.C. ABSTRACT This study was conducted to find out the important water quality factors related to eutrophication in the Te-Chi reservoir by different regression methods so that the explanation of the models could be improved, and the management plan of eutrophication can be established successfully. Three regression methods including forward, backward, and stepwise were used to analyze the important water quality factors that had been initially screened by simple regression analysis. The results showed that all standardized R2 values of three methods were better than that of the original multiple regression model and the best method was forward regression analysis. Consequently, regression methods could screen efficiently the important water quality factors of eutrophication in order to improve the explanation capabili