Contents

1 碁 螳
2 語 level 譟一
3 R
4 谿瑚襭


1 碁 螳 #

語 譬襯
  • 螻牛旧(common factor) - 蟯豸″ 蟆朱 螳
  • (unique factor)

譟郁唄
  • 覲螳 蟯蟯螻螳 伎 .
  • 豕豐 豢豢 螻 詞 螻豺襯 ろ襴 谿碁 蟲磯 伎 蟶曙企 螻褐 伎 .
  • 覈蟯 企朱 螳れ 蠍郁伎 . (KMO and Bartleet's 蟆)

碁 譬襯
  • 覿: 覲 豢 覈(譯殊焔覿, 譯殊碁, 豕磯碁 ...)
  • 語 覿: 企 螳れ企 覈語 蟆讀 覈

覦覯
  1. 覲螳 蟯襦覿 螻牛旧語 豢
  2. 豢 螻牛旧語 伎伎 覲螳 蟯蟯螻襯 る
  3. 碁(factor loading) 賊0.3 伎企 り 覺.


  • 螻牛旧瑚骸 蟯螻 伎 螳 覲 煙 螳蟆壱 襦 蠍一
  • 碁 蟆郁骸襯 覲 蟯豸° 覿襯襯 伎


factor01.png

語 伎
願唄 觜讀 譬朱 碁 る蟇 覲願 譯手朱 伎.


  • 覲伎 (譯殊焔 覿螻 觜)
  • 覲 企 譟伎 蟲譟 覦蟆
  • 語朱 覓苦伎 譴螳 覲 蟇
  • 螳 螳 豸′ 覲れ 狩 語朱 覓苦企讌
  • 蠏覿企 覲覿 る覲


  • 瑚骸 覲 伎 蟯螻襯 伎蠍 企れ 蟆曙 語螳 螳 0 1 螳蟾讌襦 碁れ 豢 る 蟆.
  • 讌螳(orthogonal rotation)
    • 碁り 蟯螻螳 襴(蟯蟯螻 0)朱 螳譯狩 (蠏覿企 覲覿 れ螻旧煙 狩蠍 伎 )
    • varimax, quartimax, equimax
  • 螳(oblique rotation)
    • 螻狩 覿覿 願姥 . 覃 碁り 蟯螻螳 襴曙 蟆 覲襦 り 覲願鍵 覓.
    • 讌螳 覦 觜 碁 讌螻, 覿 讌
    • oblimin, covarimin, quartimin, biquartimin, promax
  • R factanal varimax, promax襷 讌, 襷 rotation 讌る GPArotation れ れ覃 り .

2 語 level 譟一 #

> x <- c("5豌", "5襷")
> x <- factor(x)
> x
[1] 5豌 5襷
Levels: 5襷 5豌
> levels(x) <- c("5豌", "5襷")
> x
[1] 5襷 5豌
Levels: 5豌 5襷

3 R #

ろ 一危磯 "Excel 譟一覦覯 覦 糾覿, 誤讌, " .
tmp <- textConnection( 
  "豌蟆一  	蠍一螳	襷	豺
6	4	7	6	5
5	7	5	6	6
5	3	4	5	6
3	3	2	3	4
4	3	3	3	2
2	6	2	4	3
1	3	3	3	2
3	5	3	4	2
7	3	6	5	5
6	4	3	4	4
6	6	3	6	4
3	2	2	4	2
5	7	2	5	2
6	3	6	5	7
3	4	5	3	2
2	7	5	5	4
3	5	2	7	2
6	4	5	5	7
7	4	6	3	5
5	6	6	3	4
2	3	3	4	3
3	4	2	3	4
3	6	3	5	3
6	5	7	5	5
7	6	5	4	6") 
x <- read.table(tmp, header=TRUE) 
close.connection(tmp)
#head(x)

襾殊 譯殊焔覿螻 碁 谿企ゼ 覲伎.
#install.packages("psych")
#library("psych")
fa.parallel(x)

蟆郁骸
> fa.parallel(x)
Parallel analysis suggests that the number of factors =  2  and the number of components =  2 
語 2螳手 豢豌伎が.

facotor03.png

  • 譯殊焔 覿 5譯殊焔蟾讌 覩
  • 覿 2瑚讌襦 豢 --> 蟇碁 .

碁 企慨.
fit <- factanal(x, 2, rotation="promax") 
print(fit)
#print(fit, cutoff = 1e-05, digits = 2)

蟆郁骸
> print(fit)

Call:
factanal(x = x, factors = 2, rotation = "promax")

Uniquenesses:
豌蟆一    蠍一螳   襷     豺 
   0.339    0.869    0.419    0.005    0.287 

Loadings:
         Factor1 Factor2
豌蟆一  0.801         
            0.371 
蠍一螳  0.775         
襷            0.988 
豺      0.833         

               Factor1 Factor2
SS loadings      1.939   1.122
Proportion Var   0.388   0.224
Cumulative Var   0.388   0.612

Factor Correlations:
        Factor1 Factor2
Factor1   1.000   0.239
Factor2   0.239   1.000

Test of the hypothesis that 2 factors are sufficient.
The chi square statistic is 0.22 on 1 degree of freedom.
The p-value is 0.639 
  • Uniquenesses 煙 螳朱 旧 0.5 危企 . 豌 覲 譴 1螳 覲襯 誤螻 覈 0.5 危 螳 覲伎願 .
  • 2螳 瑚讌 . (factanal 2覯讌 襷り覲襯 3朱 覃 3螳 螻牛旧語 谿場朱朱 蟇企, 襷 覲螳 り )
  • Loadings螳 覿碁..
    • Factor1 覲企 豌蟆一, 蠍一螳, 豺 覲螳 螳 0.5伎(覲危 0.3 伎企 り 覺)朱 .
    • Factor2襯 覲企 襷 0.99襦 蟲ロ . 0.37朱 螳 .
  • Cumulative Var(る 豐覿一 螳) - Factor2蟾讌 0.612襷 る.
  • Factor Correlations - Factor1 Factor2 覲襦 蟯 .
  • p-value 0.639襦 覿 蠏覓願れ 蠍郁讌 覈詩覩襦, 碁 覩
    • 蠏覓願: 豢豢 語 豸′ 覲 伎 蟯螻襯 る(谿 )
    • 襴所: 豢豢 碁慨 襷 語 (谿 )
    • 蠏覓願れ 蠍郁覃, 碁 覓伎覩誤.

#誤 る 谿瑚 譟郁 るゴ蟆 
#fit <- factanal(x, 2, rotation="none") 
load <- fit$loadings[,1:2] 
plot(load,type="n") # set up plot 
text(load,labels=names(x),cex=.7) # add variable names 
factor02.png

factanal()螳 覃..
library(psych)
library(GPArotation)
fit <- fa(r=cor(x1), nfactors=3, rotate="promax")
summary(fit)

load <- fit$loadings[,1:2] 
plot(load,type="n") # set up plot 
text(load,labels=names(x),cex=.7) # add variable names 


螳 覃 螻牛旧瑚 蟲 .
> factanal(x, 2, rotation="promax", scores="regression")$scores
         Factor1     Factor2
 [1,]  0.8266176  1.15483124
 [2,]  0.5335642  1.24017288
 [3,]  0.5908294  0.35575418
 [4,] -0.3220569 -1.12270001
 [5,] -0.5015774 -1.07468932
 [6,] -1.0200758 -0.06511692
 [7,] -1.1099463 -0.90967064
 [8,] -0.9146350 -0.09589597
 [9,]  1.0607151  0.22680194
[10,]  0.2550042 -0.41469562
[11,] -0.1482268  1.42528154
[12,] -1.0589853 -0.06032610
[13,] -0.8896313  0.76484419
[14,]  1.4105432  0.13182575
[15,] -0.3724041 -1.10981555
[16,] -0.4344594  0.63898395
[17,] -1.6638319  2.69963963
[18,]  1.2315831  0.18258979
[19,]  1.4379888 -1.60177959
[20,]  0.7387958 -1.40917884
[21,] -0.8238109 -0.12347804
[22,] -0.3307093 -1.11890147
[23,] -0.8399416  0.74910139
[24,]  1.0109283  0.24243975
[25,]  1.3337221 -0.70601818

4 谿瑚襭 #