運用組合資料於PLS2演算法之比較 = PLS2 Algorithms ...
國立高雄大學統計學研究所

 

  • 運用組合資料於PLS2演算法之比較 = PLS2 Algorithms Comparison On Compositional Data
  • Record Type: Language materials, printed : monographic
    Paralel Title: PLS2 Algorithms Comparison On Compositional Data
    Author: 張珮甄,
    Secondary Intellectual Responsibility: 國立高雄大學
    Place of Publication: [高雄市]
    Published: 撰者;
    Year of Publication: 2012[民101]
    Description: 93面表格 : 30公分;
    Subject: 偏最小平方法
    Subject: Partial Least Squares
    Online resource: http://handle.ncl.edu.tw/11296/ndltd/59526356881451751549
    Notes: 參考書目:面26-27
    Notes: 附錄:1.表附錄;2.程式碼
    Summary: 在資料分析中,組合資料常應用在社會學、經濟學,像是投資、消費結構等領域中,因為組合資料能夠傳達資料間相對訊息,因此常用來討論變數之間的結構關係。由於組合資料被廣泛應用於各領域,PLS2的方法被廣泛應用作為分析小樣本下變數之間關係的統計工具,且在P大n小的情況下PLS2的方法近年來也被廣泛應用。本文中我們將運用組合資料於五種PLS2的方法,並介紹五種PLS2演算法則以及我們建議的修訂,最後透過模擬的方式將資料分為中心化和標準化,比較五種PLS2演算法則之間的優劣,找出能較準確預測模型的演算法則,也討論了在P大n小的情況下透過計算累積解釋變異選取較精簡的預測模型。 In data analysis, compositional data are widely used in sociology, economics, such as investment, consumption structure fields, etc. Because compositional data carry relative information, so they are often used to discuss the relationships of variables structure. The type of compositional data is widely used, and PLS2 method is widely used in discussing the relationship between variables as statistical technique in analysis under small samples. Recently, PLS2 algorithms are also discussed and applied to the issue of large P small n problem. In this paper, we will use compositional data on five PLS2 algorithms. In the end, we will compare the five PLS2 algorithms to find out whether some algorithms will give the better predicted model through simulations, and use the proportion of the total variance accounted for by the model to choose the simplified predicted model.
Items
  • 2 records • Pages 1 •
 
310002294653 博碩士論文區(二樓) 不外借資料 學位論文 TH 008M/0019 343201 1111 2012 一般使用(Normal) On shelf 0
310002294661 博碩士論文區(二樓) 不外借資料 學位論文 TH 008M/0019 343201 1111 2012 c.2 一般使用(Normal) On shelf 0
  • 2 records • Pages 1 •
Reviews
Export
pickup library
 
 
Change password
Login