透過您的圖書館登入
IP:18.216.8.40
  • 學位論文

多重誤差項的分析問題

The analysis of experimental designs involving multiple error terms

指導教授 : 鄭子韋

摘要


做資料分析時,需要考慮很多種實驗設計的方法。在不同的設計方法中做變異數分析,檢定每一項對模型是否具有影響時,根據因子為固定或是隨機,會需要用不同的檢定公式。但在傳統統計軟體中,通常因子都是對誤差均方作檢定。本論文研究在基本套件、lme4 套件、EMSaov 套件和mixlm 套件下該怎麼使用讓我們能夠得到正確的變異數分析表。 在本篇研究中,用了完全隨機設計、巢狀設計及裂區設計來做變異數分析,並用了不同的套件,分析優缺點,來取得想要得到的正確結果。 在每一章節中都有用一筆資料來分析,用不同的套件,來得到的變異數分表結果。

並列摘要


When doing data analysis, there are many experimental design methods you need to consider. In different experimental design, according to factor is fixed or random, to test whether the term in the model have any effect, we need different testing formula. But in general statistical software, testing formula’s denominator is mean squares error. How to set in stats package, lme4 package, lmerTest package, EMSaov package, and mixlm package which can directly make the correct analysis of variance table is the study of this paper. In the paper, use different packages to do analysis of variance with completely randomized design, nested design, and split plot design. And get the correct analysis of variance table. In each chapter, we use different packages and analyze the data to obtain the results of the analysis of variance table.

參考文獻


1. Bates, Douglas, Martin Mächler, Benjamin M. Bolker, and Steven C. Walker
(2015). Fitting Linear Mixed-Effects Models Using lme4. Journal of Statistical
Software, Vol.67, Issue 1.
2. Collier, Raymond O., Jr., and Frank B. Baker(1963). The Randomization
Distribution of F-Ratios for the Split plot Design–an Empirical Investigation.

延伸閱讀


國際替代計量