國立高雄大學圖資館 |

Language: English

Back

以投影追蹤進行分群分析之探索 = A Study of Clusteri...

國立高雄大學統計學研究所

以投影追蹤進行分群分析之探索 = A Study of Clustering Analysis Using Projection Pursuit

Record Type:	Language materials, printed : monographic
Paralel Title:	A Study of Clustering Analysis Using Projection Pursuit
Author:	王柏凱,
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:	69面圖，表格 : 30公分;
Subject:	投影追蹤
Subject:	projection pursuit
Online resource:	http://handle.ncl.edu.tw/11296/ndltd/76343190973663128886
Notes:	參考書目：面32-33
Summary:	資料分群分析為國內外眾多學者研究的重要課題，「投影追蹤(projection pursuit)」為資料分群的重要方法之一。根據美國學者 Friedman 與Tukey(1974) 指出「投影追蹤」在多維度資料的資料探勘分析，是將多維度資料藉線性投影轉成一維度資料或二維度資料，再進一步找出有興趣的分群結果。本研究透過 Friedman與Tukey 所使用的投影指標(P-indexes)的演算法，探索出投影指標越大並非使投影後的資料點較密集以及經投影指標值愈大的投影方向投影後的分群愈明顯。此外，當原始資料透過投影追蹤，利用更新方法去找尋到的最終投影方向投影後的資料點與原始資料透過SPSS的分群分析的結果，計算其錯誤判斷率(資料原本屬於A群，卻判斷不是A群的機率)，發現透過投影追蹤再進行分群分析其錯誤判斷率較低。 Clustering analysis is an important issue for many scholars. Projection pursuit is one of the important methods of clustering analysis. According to the American scholars, Friedman and Tukey(1974), "projection pursuit" in the exploratory data analysis of multi-dimensional data is to use multi-dimensional data converted to one dimension or two dimensions by the linear projection to find something interests in clustering results. In this study,through projection index algorithm by Friedman and Tukey,exploring the larger projection index is not more dense in projecting data and the greater projection index in projecting data can cluster more obvious. The projectional data using numerical method by the projection pursuitand the original data were used the clustering analysis in SPSS,to calculate the misjudgment rate (the probability of a datum is belonged to A, but to judge not A). Finding the data usingthe projection pursuit has lower misjudgment rate.

以投影追蹤進行分群分析之探索 = A Study of Clustering Analysis Using Projection Pursuit
王, 柏凱

以投影追蹤進行分群分析之探索 = A Study of Clustering Analysis Using Projection Pursuit / 王柏凱撰 - [高雄市] : 撰者, 2012[民101]. - 69面 ; 圖，表格 ; 30公分.
參考書目：面32-33.
投影追蹤projection pursuit

以投影追蹤進行分群分析之探索 = A Study of Clustering Analysis Using Projection Pursuit
LDR:02963nam0a2200277 450 001 360456
005 20170214090913.0
009 360456
010 0 $b 精裝
010 0 $b 平裝
100 $a 20170214d2012 k y0chiy05 e
101 1 $a chi $d chi $d eng
102 $a tw
105 $a ak am 000yy
200 1 $a 以投影追蹤進行分群分析之探索 $d A Study of Clustering Analysis Using Projection Pursuit $z eng $f 王柏凱撰
210 $a [高雄市] $c 撰者 $d 2012[民101]
215 0 $a 69面 $c 圖，表格 $d 30公分
300 $a 參考書目：面32-33
314 $a 指導教授：陳俞成博士
328 $a 碩士論文--國立高雄大學統計學研究所
330 $a 資料分群分析為國內外眾多學者研究的重要課題，「投影追蹤(projection pursuit)」為資料分群的重要方法之一。根據美國學者 Friedman 與Tukey(1974) 指出「投影追蹤」在多維度資料的資料探勘分析，是將多維度資料藉線性投影轉成一維度資料或二維度資料，再進一步找出有興趣的分群結果。本研究透過 Friedman與Tukey 所使用的投影指標(P-indexes)的演算法，探索出投影指標越大並非使投影後的資料點較密集以及經投影指標值愈大的投影方向投影後的分群愈明顯。此外，當原始資料透過投影追蹤，利用更新方法去找尋到的最終投影方向投影後的資料點與原始資料透過SPSS的分群分析的結果，計算其錯誤判斷率(資料原本屬於A群，卻判斷不是A群的機率)，發現透過投影追蹤再進行分群分析其錯誤判斷率較低。 Clustering analysis is an important issue for many scholars. Projection pursuit is one of the important methods of clustering analysis. According to the American scholars, Friedman and Tukey(1974), "projection pursuit" in the exploratory data analysis of multi-dimensional data is to use multi-dimensional data converted to one dimension or two dimensions by the linear projection to find something interests in clustering results. In this study,through projection index algorithm by Friedman and Tukey,exploring the larger projection index is not more dense in projecting data and the greater projection index in projecting data can cluster more obvious. The projectional data using numerical method by the projection pursuitand the original data were used the clustering analysis in SPSS,to calculate the misjudgment rate (the probability of a datum is belonged to A, but to judge not A). Finding the data usingthe projection pursuit has lower misjudgment rate.
510 1 $a A Study of Clustering Analysis Using Projection Pursuit $z eng
610 0 $a 投影追蹤 $a 投影指標 $a 維度縮減 $a 分群分析
610 1 $a projection pursuit $a projection index $a dimensionality reduction $a clustering analysis
681 $a 008M/0019 $b 343201 1042 $v 2007年版
700 1 $a 王 $b 柏凱 $4 撰 $3 592861
712 0 2 $a 國立高雄大學 $b 統計學研究所 $3 166081
801 0 $a tw $b NUK $c 20130306 $g CCR
856 7 $z 電子資源 $2 http $u http://handle.ncl.edu.tw/11296/ndltd/76343190973663128886