透過您的圖書館登入
IP:3.146.105.194
  • 期刊

設計矩陣與人口資料分析:線性代數之應用

Design Matrix and Data Analysis: An Application of Linear Algebra

摘要


本文引入設計矩陣的概念,使用幾個簡化的例子說明傳統的迴歸分析與變異數分析乃是同一個分析方法,也指出次數分配的分析可以參數化,適用同一套最小平方法的分析,打破傳統方法與計量教科書有關計量尺度與分析方法的迷思。本文同時指出,設計矩陣的安排對應著模型與參數設定,儘管矩陣的階序相同,不同的矩陣內容代表不同的模型與參數設定及研究設計,則進行資料分析以前,一但有了分析模型與參數設定的構想,研究者宜乎審慎考量其設計矩陣與模型設立的對應關係,確認設計矩陣正確代表其研究設計。正交的設計矩陣保證參數估計「分別」為之,互不相干,不會產生解釋變項間牽扯不清的問題,乃研究者需要多加注意的設計。在討論中,本文也指出傳統方法與計量教科書對於分析單元的主張未必能適應學術的進展,尤其在多層次分析與機率或次數分析的架構下,此一主張有需要修飾之處。

並列摘要


Introducing the concept of a ”design matrix”, this paper relates the regression analysis, the analysis of variance, and the analysis of frequency counts to the use of the least squares method. It is demonstrated with some simplified examples showing that, given a properly arranged design matrix, the analyses of means, probabilities, and frequencies share the same least squares approach as that of the regression analysis. Special attention is called to the ”orthogonal” design matrix which results in independent parameter estimations jointly. Researchers are urged to review the possible arrangements of the design matrix before actually doing on the estimation.

參考文獻


Fisher,Franklin.(1970).Tests of Equality between Sets of Coefficients in- Two Linear Regressions: An Expository Notes.Econometrica.38,361-366.
Goodman, L.(1984).The Analysis of Cross-Classified Data Having Ordered Categories.Cambridge:Harvard University Press.
Iversen,Gudmund R.,Helmut Norpoth.(1987).Analysis of Variance.Thousand Oaks, CA:Sage Publications, Inc.
McCullagh, P.,J. A. Nelder.,Second(Ed)(1999).Generalized Linear Models.Boca Raton, FL:Chapman & Hall, CRC Press.
Raudenbush, S. W.,A. S. Bryk.(2002).Hierarchical Linear Models: Applications and Data Analysis Methods.Thousand Oaks, CA:Sage Publications, Inc.

被引用紀錄


Hsieh, H. T. (2011). 具週期性前導序列之正交分頻多工/正交分頻多工存取系統中最大似然載波頻率偏移估計 [doctoral dissertation, National Chiao Tung University]. Airiti Library. https://doi.org/10.6842/NCTU.2011.00268

延伸閱讀