Please use this identifier to cite or link to this item:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/19520
Full metadata record
???org.dspace.app.webui.jsptag.ItemTag.dcfield??? | Value | Language |
---|---|---|
dc.contributor.advisor | 陳正剛 | |
dc.contributor.author | Shang-Feng Chen | en |
dc.contributor.author | 陳上峰 | zh_TW |
dc.date.accessioned | 2021-06-08T02:03:12Z | - |
dc.date.copyright | 2016-04-15 | |
dc.date.issued | 2016 | |
dc.date.submitted | 2016-03-31 | |
dc.identifier.citation | [1] J.W.Johnson,&J.M. Lebreton ,“History and use of relative importance indices in organizational research”,Organizational Research Methods,2004,7,238-257.
[2] Budescu, D.V. ,1993. Dominance Analysis: A new approach to the problem of relative importance of predictors in multiple regression. Psychological Bulletin, 114, 542-551C. D. Jones, A. B. Smith, and E.F. Roberts, Book Title, Publisher, Location, Date. [3] J.W.Johnson,“ A Heuristic Method for Estimating the Relative Weight of Predictor Variables in Multiple Regression”,Multivariate Behavioral Research,2000,35,1-19. [4] Azen, R., & Budescu, D.V.,“ Comparing predictors in multivariate regression models: An extension of dominance analysis”,Journal of Behavioral and Educational Statistics, 2006, 31, 157-180. [5] R.M. Johnson, “The minimal transformation to orthonormality”, Psychomatrika,1966,31, 61-66. [6] Lebreton, J.M. , Ployhart, R.E. , and Ladd, R.T. ,2004. A Monte Carlo comparison of relative importance methodologies. Organizational Research Methods, 7, 258-282 [7] Chao, Y. E. , Zhao, Y. , Kupper, L. L. , and Nylander-French, L.A., 2008. Quantifying the relative importance of predictors in multiple linear regression analyses for public health studies. Journal of Occupational and Environmental Hygiene, 5(8), 519 - 529. [8] C. Radhakrishna Rao., 1981. A lemma on G-inverse of a matrix and computation of correlation coefficients in the singular case. Communications in Statistics – Theory and Methods, 10:1, 1- 10. [9] Jeff W. Johnson, Scott Tonidandel, and James M. Lebreton, Determining the statistical significance of relative weight. Psychological Methods, 2009, 14,, 387 - 399. [10] William Kruskal, Concepts of relative importance, Psychometrika,, 1984, 39 – 45. [11] Johan Bring, A geometric approach to compare variables in a regression model, The American Statistician, 1996, 50, 57-62. [12] Ledyard R Tucker, Lee G. Cooper, and William Meredith, Obtaining squared multiple correlations from a correlation matrix which may be singular, Psychometrika, 1972, 37, 143 – 148. | |
dc.identifier.uri | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/19520 | - |
dc.description.abstract | 變數重要性 ( Predictor Importance ) 指的是在複迴歸模型中,個別獨立變數對相依變數所發揮的解釋能力,而根據Johnson(2004)對變數重要性的定義,變數重要性同時須考慮獨立變數獨自對相依變數的解釋能力,以及在模型中與其他獨立變數共同解釋時所發揮的解釋能力,然而後者的效果由於獨立變數間往往存在複雜的共線性問題,所以使得變數重要性無法明確地衡量。
在變數重要性的相關文獻中已提出多種評斷變數重要性的方法,但能對共線性問題提出合理解釋與解決方法的目前有兩個主要指標,分別是Budescu(1993)的Dominance Index與Johoson(2000)的Relative Weight,這兩個指標已經經過許多文獻的探討與實例的分析,尤其Dominance Index發展較早,具有明確的解釋意義,被視為變數重要性研究最具參考性指標。 然而目前變數重要性的相關研究,都是樣本數充足的前提下進行討論,也就是假設獨立變數間為線性獨立 ( linear independent ) ,因此本研究的主要目的為分析Dominance Index與Relative Weight在獨立變數間為線性相依下 ( linear dependent ),是否還能夠合理的反映變數重要性,而本文首先將模擬出含有三個獨立變數的線性相依案例,並說明在此模擬案例中合理的變數重要性為何,接著指出幾個Dominance Index 與Relative Weight在獨立變數為線性相依時的不合理之處。 基於Dominance Index與Relative Weight在線性相依案例下皆無法合理反映出變數重要性,本文提出變數加權重要性 ( Predictor Weighted Importance ) 來改進原先變數重要性指標在線性相依案例下之表現,主要概念是將原先之線性相依案例,轉化為多個線性獨立的sub case,計算出各sub case的變數重要性後,再決定各個sub case的計算權重,將各sub case之變數重要性進行加權總合得到變數加權重要性,本文將以此方法建立Weighted Dominance Index與Weighted Relative Weight,並以較複雜之模擬案例及實際案例來進行變數重要性以及使用採用不同模型權重之變數加權重要性的效果比較。 | zh_TW |
dc.description.provenance | Made available in DSpace on 2021-06-08T02:03:12Z (GMT). No. of bitstreams: 1 ntu-105-R02546033-1.pdf: 3895922 bytes, checksum: debf4935c6a9d4c02008a5d94f708910 (MD5) Previous issue date: 2016 | en |
dc.description.tableofcontents | 口試委員會審定書 #
誌謝 i 中文摘要 ii Abstract iii CONTENTS iv LIST OF FIGURES vi LIST OF TABLES vii Chapter 1 緒論 1 1.1 研究背景 1 1.2 研究動機與目標 3 Chapter 2 文獻探討 5 2.1 Dominance Index 5 2.2 Relative Weight 7 2.2.1 資料矩陣之最佳正交近似轉換方法 7 2.2.2 原始變數X之相對重要性 9 2.3 概括性相對重要性指標 10 2.3.1 非滿秩資料矩陣之最佳正交近似轉換方法 10 2.3.2 非滿秩最佳正交轉換變數之變數重要性 12 2.3.3 非滿秩下相對重要性指標之建構 13 Chapter 3 非滿秩之Dominance Index 改善方法 15 3.1 非滿秩下Dominance Index之不合理處 15 3.1.1 變數重要性之間的比例 17 3.1.2 變數重要性的變動趨勢 19 3.1.3 變數重要性之變動趨勢的連續性 24 3.2 重新定義子模型計算權重之方法 25 3.3 模型權重之計算方法 30 3.3.1 以sub case內獨立變數之共線性作為加權依據 30 3.3.2 以被刪去獨立變數與sub case內獨立變數之平方相關係數作為加權依據 30 3.3.3 以sub case內獨立變數之平方標準迴歸係數作為加權依據 31 3.4 不同加權方法之效果比較 32 Chapter 4 加權方法應用於Relative Weight 37 4.1 將完整模型轉換成多個sub case之方法 37 4.2 模型權重之計算方法 42 4.2.1 以最大距離權重獨立變數與其餘獨立變數之平方相關係數作為加權依據 42 4.2.2 各sub case模型權重相等 43 4.3 不同模型權重之效果比較 47 Chapter 5 案例分析 50 5.1 模擬案例分析 50 5.1.1 Weighted Dominance Index 50 5.1.2 Weighted Relative Weight 53 5.2 實際案例分析 55 Chapter 6 結論 59 參考文獻 60 | |
dc.language.iso | zh-TW | |
dc.title | 非滿秩情形下之預測變數加權重要性研究 | zh_TW |
dc.title | Study of Predictor Weighted Importance under Low-Rank Condition | en |
dc.type | Thesis | |
dc.date.schoolyear | 104-2 | |
dc.description.degree | 碩士 | |
dc.contributor.oralexamcommittee | 藍俊宏,洪弘 | |
dc.subject.keyword | 變數重要性,Dominance Index,Relative Weight,預測變數加權重要性, | zh_TW |
dc.subject.keyword | Predictor Importance,Dominance Index,Relative Weight,Weighted Predictor Importance, | en |
dc.relation.page | 61 | |
dc.identifier.doi | 10.6342/NTU201600173 | |
dc.rights.note | 未授權 | |
dc.date.accepted | 2016-03-31 | |
dc.contributor.author-college | 工學院 | zh_TW |
dc.contributor.author-dept | 工業工程學研究所 | zh_TW |
Appears in Collections: | 工業工程學研究所 |
Files in This Item:
File | Size | Format | |
---|---|---|---|
ntu-105-1.pdf Restricted Access | 3.8 MB | Adobe PDF |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.