第6回（12月3日） Task Check and Weekly Assignment
To Do
う・ん・ち・く
□ 因子分析法Factor Analysisとは心理学における統計技法の利用の中では伝統的な技術の一つで，潜在構造を明らかにする探索的因子分析をへて構造方程式モデリング（確認的因子分析）につなげると当てはめやすいのだ！
Using Package; lavaan,psych
Assignment
■ 投手のデータ(base_pitch.csv)をつかって探索的因子分析を行い，構造方程式モデリングで確認してみてください。
まずはデータセットを読み込んでください。これはプロ野球選手のデータで， 2013年の成績，年俸などが含まれているデータセットです。出典はこちら。＞＞http://baseball-data.com/ ついでに標準化しておきましょうね。
batter <- read.csv("base_bat.csv", na.strings = "NA")
summary(batter)
## money year age hight
## Min. : 8600 Min. : 1.00 Min. :24.0 Min. :169
## 1st Qu.:11375 1st Qu.: 7.75 1st Qu.:30.0 1st Qu.:177
## Median :16000 Median :12.00 Median :32.5 Median :180
## Mean :17488 Mean :11.12 Mean :33.4 Mean :181
## 3rd Qu.:20000 3rd Qu.:15.25 3rd Qu.:37.0 3rd Qu.:184
## Max. :57000 Max. :25.00 Max. :42.0 Max. :196
## NA's :5 NA's :5 NA's :5 NA's :5
## weight daritu shiai daseki dasuu
## Min. : 67.0 Min. :0.185 Min. : 9 Min. : 10 Min. : 10
## 1st Qu.: 80.8 1st Qu.:0.230 1st Qu.: 69 1st Qu.:238 1st Qu.:212
## Median : 87.5 Median :0.266 Median :108 Median :424 Median :374
## Mean : 88.4 Mean :0.262 Mean :102 Mean :395 Mean :347
## 3rd Qu.: 94.0 3rd Qu.:0.290 3rd Qu.:141 3rd Qu.:581 3rd Qu.:508
## Max. :130.0 Max. :0.333 Max. :144 Max. :658 Max. :590
## NA's :5
## anda HR datgen steal
## Min. : 2 Min. : 0 Min. : 1.0 Min. : 0.00
## 1st Qu.: 50 1st Qu.: 2 1st Qu.: 17.0 1st Qu.: 0.00
## Median : 88 Median : 6 Median : 40.0 Median : 1.00
## Mean : 95 Mean :10 Mean : 45.7 Mean : 5.65
## 3rd Qu.:147 3rd Qu.:16 3rd Qu.: 67.0 3rd Qu.: 6.00
## Max. :180 Max. :41 Max. :136.0 Max. :47.00
##
## four dead K gida
## Min. : 0.0 Min. : 0.00 Min. : 3.0 Min. : 0.00
## 1st Qu.: 16.0 1st Qu.: 1.00 1st Qu.: 38.0 1st Qu.: 0.00
## Median : 33.0 Median : 2.00 Median : 54.0 Median : 1.00
## Mean : 38.8 Mean : 3.63 Mean : 62.6 Mean : 3.19
## 3rd Qu.: 54.0 3rd Qu.: 5.00 3rd Qu.: 81.0 3rd Qu.: 5.00
## Max. :105.0 Max. :15.00 Max. :164.0 Max. :19.00
##
batter.z <- as.data.frame(scale(batter))
今回必要なパッケージ，lavaanとpsychを装着します（持ってない人はインストールしてください）
# install.packages('psych')
library(lavaan)
## This is lavaan 0.5-15
## lavaan is BETA software! Please report any bugs.
library(psych)
データセットの中でつかう変数だけ取り出しておきます。
batter2 <- batter[6:16]
想定する因子（潜在変数）の数を決めます。 決める方法はいろいろありますが，今回は平行分析というのをやってみます。
fa.parallel(batter2)
## Loading required package: MASS
## Parallel analysis suggests that the number of factors = 1 and the number of components = 1
2つのfactorを設定するのがいいよ，ということなので，そうした因子分析を考えます。
fa.result <- fa(batter2, nfactors = 2, fm = "minres", rotate = "promax")
print(fa.result, sort = T, digit = 3, cut = 0.3)
## Factor Analysis using method = minres
## Call: fa(r = batter2, nfactors = 2, rotate = "promax", fm = "minres")
## Standardized loadings (pattern matrix) based upon correlation matrix
## item MR1 MR2 h2 u2 com
## anda 5 0.894 0.981 0.01904 1.03
## dasuu 4 0.892 0.993 0.00716 1.24
## daseki 3 0.800 0.992 0.00814 1.19
## shiai 2 0.750 0.922 0.07767 1.05
## daritu 1 0.646 0.495 0.50457 1.04
## steal 8 0.607 0.262 0.73808 1.00
## four 9 0.940 0.901 0.09852 1.52
## HR 6 0.834 0.688 0.31171 1.11
## dead 10 0.645 0.439 0.56068 1.00
## datgen 7 0.340 0.639 0.844 0.15606 1.00
## K 11 0.344 0.522 0.656 0.34394 1.73
##
## MR1 MR2
## SS loadings 4.563 3.612
## Proportion Var 0.415 0.328
## Cumulative Var 0.415 0.743
## Proportion Explained 0.558 0.442
## Cumulative Proportion 0.558 1.000
##
## With factor correlations of
## MR1 MR2
## MR1 1.000 0.738
## MR2 0.738 1.000
##
## Mean item complexity = 1.2
## Test of the hypothesis that 2 factors are sufficient.
##
## The degrees of freedom for the null model are 55 and the objective function was 23.78 with Chi Square of 1225
## The degrees of freedom for the model are 34 and the objective function was 7.844
##
## The root mean square of the residuals (RMSR) is 0.065
## The df corrected root mean square of the residuals is 0.117
##
## The harmonic number of observations is 57 with the empirical chi square 26.67 with prob < 0.811
## The total number of observations was 57 with MLE Chi Square = 393.5 with prob < 9.39e-63
##
## Tucker Lewis Index of factoring reliability = 0.489
## RMSEA index = 0.4613 and the 90 % confidence intervals are 0.393 0.4694
## BIC = 256
## Fit based upon off diagonal values = 0.991
## Measures of factor score adequacy
## MR1 MR2
## Correlation of scores with factors 0.999 0.993
## Multiple R square of scores with factors 0.997 0.987
## Minimum correlation of possible factor scores 0.994 0.974
結果として，どこがどのようにまとまるのかがわかりましたので，それにあわせて方程式モデリングをします。
model <- "
jibun =~ anda+dasuu+daseki+shiai+daritu+steal+datgen+K
aite =~ four+HR+dead+datgen+K
jibun ~~ aite
money ~jibun + aite
"
sem.result <- sem(model, data = batter.z)
summary(sem.result, fit.measures = TRUE, standardized = T)
## lavaan (0.5-15) converged normally after 136 iterations
##
## Used Total
## Number of observations 52 57
##
## Estimator ML
## Minimum Function Test Statistic 366.564
## Degrees of freedom 50
## P-value (Chi-square) 0.000
##
## Model test baseline model:
##
## Minimum Function Test Statistic 1239.640
## Degrees of freedom 66
## P-value 0.000
##
## User model versus baseline model:
##
## Comparative Fit Index (CFI) 0.730
## Tucker-Lewis Index (TLI) 0.644
##
## Loglikelihood and Information Criteria:
##
## Loglikelihood user model (H0) -433.988
## Loglikelihood unrestricted model (H1) -250.706
##
## Number of free parameters 28
## Akaike (AIC) 923.976
## Bayesian (BIC) 978.611
## Sample-size adjusted Bayesian (BIC) 890.683
##
## Root Mean Square Error of Approximation:
##
## RMSEA 0.349
## 90 Percent Confidence Interval 0.316 0.383
## P-value RMSEA <= 0.05 0.000
##
## Standardized Root Mean Square Residual:
##
## SRMR 0.104
##
## Parameter estimates:
##
## Information Expected
## Standard Errors Standard
##
## Estimate Std.err Z-value P(>|z|) Std.lv Std.all
## Latent variables:
## jibun =~
## anda 1.000 0.937 0.979
## dasuu 0.999 0.030 32.996 0.000 0.936 0.998
## daseki 0.997 0.031 32.086 0.000 0.934 0.997
## shiai 0.946 0.047 19.986 0.000 0.886 0.960
## daritu 0.715 0.111 6.436 0.000 0.670 0.673
## steal 0.518 0.134 3.867 0.000 0.486 0.475
## datgen 0.342 0.057 6.007 0.000 0.321 0.324
## K 0.540 0.127 4.252 0.000 0.506 0.518
## aite =~
## four 1.000 0.655 0.674
## HR 1.504 0.238 6.317 0.000 0.986 0.973
## dead 0.981 0.225 4.369 0.000 0.643 0.639
## datgen 1.111 0.187 5.932 0.000 0.728 0.735
## K 0.498 0.197 2.524 0.012 0.326 0.334
##
## Regressions:
## money ~
## jibun -0.205 0.189 -1.084 0.278 -0.192 -0.194
## aite 0.770 0.297 2.591 0.010 0.504 0.509
##
## Covariances:
## jibun ~~
## aite 0.422 0.124 3.414 0.001 0.688 0.688
##
## Variances:
## anda 0.038 0.008 0.038 0.042
## dasuu 0.003 0.002 0.003 0.004
## daseki 0.006 0.002 0.006 0.006
## shiai 0.068 0.014 0.068 0.079
## daritu 0.542 0.107 0.542 0.547
## steal 0.808 0.159 0.808 0.774
## datgen 0.026 0.018 0.026 0.027
## K 0.366 0.072 0.366 0.383
## four 0.517 0.104 0.517 0.546
## HR 0.054 0.032 0.054 0.053
## dead 0.601 0.120 0.601 0.592
## money 0.823 0.163 0.823 0.839
## jibun 0.878 0.180 1.000 1.000
## aite 0.429 0.157 1.000 1.000
で，ちょっと図で確認してみたりして。
# install.packages('semPlot')
library(semPlot)
## This is semPlot 0.3.3
## semPlot is BETA software! Please report any bugs.
semPaths(sem.result, "std")
ここからモデルを改良していった方が楽よね。
modificationindices(sem.result)
## lhs op rhs mi epc sepc.lv sepc.all sepc.nox
## 1 jibun =~ anda NA NA NA NA NA
## 2 jibun =~ dasuu 0.000 0.000 0.000 0.000 0.000
## 3 jibun =~ daseki 0.000 0.000 0.000 0.000 0.000
## 4 jibun =~ shiai 0.000 0.000 0.000 0.000 0.000
## 5 jibun =~ daritu 0.000 0.000 0.000 0.000 0.000
## 6 jibun =~ steal 0.000 0.000 0.000 0.000 0.000
## 7 jibun =~ datgen 0.000 0.000 0.000 0.000 0.000
## 8 jibun =~ K 0.000 0.000 0.000 0.000 0.000
## 9 jibun =~ four 16.078 0.612 0.573 0.589 0.589
## 10 jibun =~ HR 13.705 -0.642 -0.602 -0.594 -0.594
## 11 jibun =~ dead 0.939 0.159 0.149 0.148 0.148
## 12 aite =~ anda 4.008 0.120 0.079 0.082 0.082
## 13 aite =~ dasuu 8.024 -0.076 -0.050 -0.053 -0.053
## 14 aite =~ daseki 3.135 0.050 0.033 0.035 0.035
## 15 aite =~ shiai 0.520 0.057 0.037 0.040 0.040
## 16 aite =~ daritu 3.292 0.400 0.262 0.263 0.263
## 17 aite =~ steal 9.418 -0.825 -0.540 -0.529 -0.529
## 18 aite =~ datgen 0.000 0.000 0.000 0.000 0.000
## 19 aite =~ K 0.000 0.000 0.000 0.000 0.000
## 20 aite =~ four NA NA NA NA NA
## 21 aite =~ HR 0.000 0.000 0.000 0.000 0.000
## 22 aite =~ dead 0.000 0.000 0.000 0.000 0.000
## 23 anda ~~ anda 0.000 0.000 0.000 0.000 0.000
## 24 anda ~~ dasuu 4.333 0.008 0.008 0.009 0.009
## 25 anda ~~ daseki 6.599 -0.010 -0.010 -0.011 -0.011
## 26 anda ~~ shiai 3.541 -0.014 -0.014 -0.016 -0.016
## 27 anda ~~ daritu 36.987 0.125 0.125 0.131 0.131
## 28 anda ~~ steal 0.038 0.005 0.005 0.005 0.005
## 29 anda ~~ datgen 11.048 0.021 0.021 0.022 0.022
## 30 anda ~~ K 10.546 -0.055 -0.055 -0.059 -0.059
## 31 anda ~~ four 8.871 -0.060 -0.060 -0.065 -0.065
## 32 anda ~~ HR 0.784 -0.008 -0.008 -0.008 -0.008
## 33 anda ~~ dead 0.010 -0.002 -0.002 -0.002 -0.002
## 34 dasuu ~~ dasuu 0.000 0.000 0.000 0.000 0.000
## 35 dasuu ~~ daseki 0.427 0.004 0.004 0.005 0.005
## 36 dasuu ~~ shiai 1.965 -0.005 -0.005 -0.006 -0.006
## 37 dasuu ~~ daritu 2.771 -0.015 -0.015 -0.016 -0.016
## 38 dasuu ~~ steal 0.003 -0.001 -0.001 -0.001 -0.001
## 39 dasuu ~~ datgen 0.011 0.000 0.000 0.000 0.000
## 40 dasuu ~~ K 0.002 0.000 0.000 0.000 0.000
## 41 dasuu ~~ four 11.130 -0.030 -0.030 -0.032 -0.032
## 42 dasuu ~~ HR 0.331 0.002 0.002 0.002 0.002
## 43 dasuu ~~ dead 5.917 -0.023 -0.023 -0.025 -0.025
## 44 daseki ~~ daseki 0.000 0.000 0.000 0.000 0.000
## 45 daseki ~~ shiai 6.521 0.010 0.010 0.012 0.012
## 46 daseki ~~ daritu 1.286 -0.011 -0.011 -0.012 -0.012
## 47 daseki ~~ steal 0.681 0.010 0.010 0.010 0.010
## 48 daseki ~~ datgen 2.644 -0.005 -0.005 -0.005 -0.005
## 49 daseki ~~ K 2.199 0.012 0.012 0.013 0.013
## 50 daseki ~~ four 26.661 0.048 0.048 0.053 0.053
## 51 daseki ~~ HR 0.213 -0.002 -0.002 -0.002 -0.002
## 52 daseki ~~ dead 6.683 0.026 0.026 0.027 0.027
## 53 shiai ~~ shiai 0.000 0.000 0.000 0.000 0.000
## 54 shiai ~~ daritu 0.576 -0.020 -0.020 -0.022 -0.022
## 55 shiai ~~ steal 2.017 -0.047 -0.047 -0.049 -0.049
## 56 shiai ~~ datgen 0.108 -0.003 -0.003 -0.003 -0.003
## 57 shiai ~~ K 0.186 0.010 0.010 0.011 0.011
## 58 shiai ~~ four 1.291 0.030 0.030 0.034 0.034
## 59 shiai ~~ HR 0.002 0.001 0.001 0.001 0.001
## 60 shiai ~~ dead 0.019 0.004 0.004 0.004 0.004
## 61 daritu ~~ daritu 0.000 0.000 0.000 0.000 0.000
## 62 daritu ~~ steal 0.002 -0.004 -0.004 -0.004 -0.004
## 63 daritu ~~ datgen 5.462 0.054 0.054 0.055 0.055
## 64 daritu ~~ K 6.664 -0.160 -0.160 -0.165 -0.165
## 65 daritu ~~ four 2.871 -0.126 -0.126 -0.130 -0.130
## 66 daritu ~~ HR 0.285 -0.017 -0.017 -0.017 -0.017
## 67 daritu ~~ dead 0.083 0.023 0.023 0.023 0.023
## 68 steal ~~ steal 0.000 0.000 0.000 0.000 0.000
## 69 steal ~~ datgen 4.493 -0.060 -0.060 -0.059 -0.059
## 70 steal ~~ K 4.784 0.166 0.166 0.166 0.166
## 71 steal ~~ four 5.447 0.211 0.211 0.213 0.213
## 72 steal ~~ HR 0.108 -0.013 -0.013 -0.012 -0.012
## 73 steal ~~ dead 1.760 0.129 0.129 0.126 0.126
## 74 datgen ~~ datgen 0.000 0.000 0.000 0.000 0.000
## 75 datgen ~~ K 0.640 -0.017 -0.017 -0.017 -0.017
## 76 datgen ~~ four 2.048 -0.041 -0.041 -0.042 -0.042
## 77 datgen ~~ HR 18.009 0.263 0.263 0.263 0.263
## 78 datgen ~~ dead 0.054 -0.007 -0.007 -0.007 -0.007
## 79 K ~~ K 0.000 0.000 0.000 0.000 0.000
## 80 K ~~ four 1.013 0.062 0.062 0.065 0.065
## 81 K ~~ HR 0.535 0.021 0.021 0.021 0.021
## 82 K ~~ dead 0.294 0.036 0.036 0.036 0.036
## 83 four ~~ four 0.000 0.000 0.000 0.000 0.000
## 84 four ~~ HR 1.394 -0.047 -0.047 -0.048 -0.048
## 85 four ~~ dead 2.415 0.123 0.123 0.125 0.125
## 86 HR ~~ HR 0.000 0.000 0.000 0.000 0.000
## 87 HR ~~ dead 0.916 -0.039 -0.039 -0.038 -0.038
## 88 dead ~~ dead 0.000 0.000 0.000 0.000 0.000
## 89 money ~~ money 0.000 0.000 0.000 0.000 0.000
## 90 jibun ~~ jibun 0.000 0.000 0.000 0.000 0.000
## 91 jibun ~~ aite 0.000 0.000 0.000 0.000 0.000
## 92 aite ~~ aite 0.000 0.000 0.000 0.000 0.000
## 93 money ~ jibun 0.000 0.000 0.000 0.000 0.000
## 94 money ~ aite 0.000 0.000 0.000 0.000 0.000
## 95 jibun ~ money NA NA NA NA NA
## 96 jibun ~ aite NA NA NA NA NA
## 97 aite ~ money NA NA NA NA NA
## 98 aite ~ jibun NA NA NA NA NA
model.2 <- "
jibun =~ anda+dasuu+daseki+shiai+daritu+steal+datgen+K
aite =~ four+HR+dead+datgen+K
jibun ~~ aite
anda ~~ daritu
daseki ~~ four
money ~ aite
"
sem.result.2 <- sem(model.2, data = batter.z)
summary(sem.result.2, fit.measures = TRUE, standardized = T)
## lavaan (0.5-15) converged normally after 148 iterations
##
## Used Total
## Number of observations 52 57
##
## Estimator ML
## Minimum Function Test Statistic 209.031
## Degrees of freedom 49
## P-value (Chi-square) 0.000
##
## Model test baseline model:
##
## Minimum Function Test Statistic 1239.640
## Degrees of freedom 66
## P-value 0.000
##
## User model versus baseline model:
##
## Comparative Fit Index (CFI) 0.864
## Tucker-Lewis Index (TLI) 0.816
##
## Loglikelihood and Information Criteria:
##
## Loglikelihood user model (H0) -355.222
## Loglikelihood unrestricted model (H1) -250.706
##
## Number of free parameters 29
## Akaike (AIC) 768.444
## Bayesian (BIC) 825.030
## Sample-size adjusted Bayesian (BIC) 733.961
##
## Root Mean Square Error of Approximation:
##
## RMSEA 0.251
## 90 Percent Confidence Interval 0.216 0.286
## P-value RMSEA <= 0.05 0.000
##
## Standardized Root Mean Square Residual:
##
## SRMR 0.228
##
## Parameter estimates:
##
## Information Expected
## Standard Errors Standard
##
## Estimate Std.err Z-value P(>|z|) Std.lv Std.all
## Latent variables:
## jibun =~
## anda 1.000 0.937 0.979
## dasuu 1.001 0.029 34.429 0.000 0.938 1.000
## daseki 0.900 0.027 33.898 0.000 0.843 0.989
## shiai 0.942 0.048 19.511 0.000 0.883 0.956
## daritu 0.709 0.093 7.634 0.000 0.664 0.667
## steal 0.517 0.134 3.855 0.000 0.484 0.474
## datgen 0.394 0.055 7.221 0.000 0.369 0.373
## K 0.550 0.120 4.584 0.000 0.515 0.527
## aite =~
## four 1.000 0.093 0.101
## HR 10.836 4.648 2.331 0.020 1.003 0.990
## dead 6.723 3.101 2.168 0.030 0.622 0.618
## datgen 7.444 3.196 2.329 0.020 0.689 0.696
## K 3.537 1.933 1.830 0.067 0.327 0.335
##
## Regressions:
## money ~
## aite 4.062 2.214 1.834 0.067 0.376 0.380
##
## Covariances:
## jibun ~~
## aite 0.056 0.028 1.990 0.047 0.646 0.646
## anda ~~
## daritu 0.124 0.026 4.674 0.000 0.124 0.855
## daseki ~~
## four 0.115 0.023 5.025 0.000 0.115 0.980
##
## Variances:
## anda 0.038 0.008 0.038 0.042
## dasuu 0.000 0.001 0.000 0.000
## daseki 0.016 0.003 0.016 0.023
## shiai 0.073 0.014 0.073 0.085
## daritu 0.550 0.108 0.550 0.555
## steal 0.810 0.159 0.810 0.776
## datgen 0.041 0.020 0.041 0.041
## K 0.365 0.072 0.365 0.383
## four 0.839 0.164 0.839 0.990
## HR 0.020 0.040 0.020 0.019
## dead 0.627 0.124 0.627 0.618
## money 0.839 0.165 0.839 0.856
## jibun 0.878 0.180 1.000 1.000
## aite 0.009 0.007 1.000 1.000
追記） 修正指数が出過ぎて困るというあなたへ。modificationindeces関数はdata.frameで返すから，受け取ってsortすれば良い。トップ5だけ欲しい，といった場合はhead関数を使えば(デフォルトで6位まで)出る。
modi <- modindices(sem.result)
modi[order(modi$mi, decreasing = T), ]
## lhs op rhs mi epc sepc.lv sepc.all sepc.nox
## 1 anda ~~ daritu 36.987 0.125 0.125 0.131 0.131
## 2 daseki ~~ four 26.661 0.048 0.048 0.053 0.053
## 3 datgen ~~ HR 18.009 0.263 0.263 0.263 0.263
## 4 jibun =~ four 16.078 0.612 0.573 0.589 0.589
## 5 jibun =~ HR 13.705 -0.642 -0.602 -0.594 -0.594
## 6 dasuu ~~ four 11.130 -0.030 -0.030 -0.032 -0.032
## 7 anda ~~ datgen 11.048 0.021 0.021 0.022 0.022
## 8 anda ~~ K 10.546 -0.055 -0.055 -0.059 -0.059
## 9 aite =~ steal 9.418 -0.825 -0.540 -0.529 -0.529
## 10 anda ~~ four 8.871 -0.060 -0.060 -0.065 -0.065
## 11 aite =~ dasuu 8.024 -0.076 -0.050 -0.053 -0.053
## 12 daseki ~~ dead 6.683 0.026 0.026 0.027 0.027
## 13 daritu ~~ K 6.664 -0.160 -0.160 -0.165 -0.165
## 14 anda ~~ daseki 6.599 -0.010 -0.010 -0.011 -0.011
## 15 daseki ~~ shiai 6.521 0.010 0.010 0.012 0.012
## 16 dasuu ~~ dead 5.917 -0.023 -0.023 -0.025 -0.025
## 17 daritu ~~ datgen 5.462 0.054 0.054 0.055 0.055
## 18 steal ~~ four 5.447 0.211 0.211 0.213 0.213
## 19 steal ~~ K 4.784 0.166 0.166 0.166 0.166
## 20 steal ~~ datgen 4.493 -0.060 -0.060 -0.059 -0.059
## 21 anda ~~ dasuu 4.333 0.008 0.008 0.009 0.009
## 22 aite =~ anda 4.008 0.120 0.079 0.082 0.082
## 23 anda ~~ shiai 3.541 -0.014 -0.014 -0.016 -0.016
## 24 aite =~ daritu 3.292 0.400 0.262 0.263 0.263
## 25 aite =~ daseki 3.135 0.050 0.033 0.035 0.035
## 26 daritu ~~ four 2.871 -0.126 -0.126 -0.130 -0.130
## 27 dasuu ~~ daritu 2.771 -0.015 -0.015 -0.016 -0.016
## 28 daseki ~~ datgen 2.644 -0.005 -0.005 -0.005 -0.005
## 29 four ~~ dead 2.415 0.123 0.123 0.125 0.125
## 30 daseki ~~ K 2.199 0.012 0.012 0.013 0.013
## 31 datgen ~~ four 2.048 -0.041 -0.041 -0.042 -0.042
## 32 shiai ~~ steal 2.017 -0.047 -0.047 -0.049 -0.049
## 33 dasuu ~~ shiai 1.965 -0.005 -0.005 -0.006 -0.006
## 34 steal ~~ dead 1.760 0.129 0.129 0.126 0.126
## 35 four ~~ HR 1.394 -0.047 -0.047 -0.048 -0.048
## 36 shiai ~~ four 1.291 0.030 0.030 0.034 0.034
## 37 daseki ~~ daritu 1.286 -0.011 -0.011 -0.012 -0.012
## 38 K ~~ four 1.013 0.062 0.062 0.065 0.065
## 39 jibun =~ dead 0.939 0.159 0.149 0.148 0.148
## 40 HR ~~ dead 0.916 -0.039 -0.039 -0.038 -0.038
## 41 anda ~~ HR 0.784 -0.008 -0.008 -0.008 -0.008
## 42 daseki ~~ steal 0.681 0.010 0.010 0.010 0.010
## 43 datgen ~~ K 0.640 -0.017 -0.017 -0.017 -0.017
## 44 shiai ~~ daritu 0.576 -0.020 -0.020 -0.022 -0.022
## 45 K ~~ HR 0.535 0.021 0.021 0.021 0.021
## 46 aite =~ shiai 0.520 0.057 0.037 0.040 0.040
## 47 dasuu ~~ daseki 0.427 0.004 0.004 0.005 0.005
## 48 dasuu ~~ HR 0.331 0.002 0.002 0.002 0.002
## 49 K ~~ dead 0.294 0.036 0.036 0.036 0.036
## 50 daritu ~~ HR 0.285 -0.017 -0.017 -0.017 -0.017
## 51 daseki ~~ HR 0.213 -0.002 -0.002 -0.002 -0.002
## 52 shiai ~~ K 0.186 0.010 0.010 0.011 0.011
## 53 shiai ~~ datgen 0.108 -0.003 -0.003 -0.003 -0.003
## 54 steal ~~ HR 0.108 -0.013 -0.013 -0.012 -0.012
## 55 daritu ~~ dead 0.083 0.023 0.023 0.023 0.023
## 56 datgen ~~ dead 0.054 -0.007 -0.007 -0.007 -0.007
## 57 anda ~~ steal 0.038 0.005 0.005 0.005 0.005
## 58 shiai ~~ dead 0.019 0.004 0.004 0.004 0.004
## 59 dasuu ~~ datgen 0.011 0.000 0.000 0.000 0.000
## 60 anda ~~ dead 0.010 -0.002 -0.002 -0.002 -0.002
## 61 dasuu ~~ steal 0.003 -0.001 -0.001 -0.001 -0.001
## 62 dasuu ~~ K 0.002 0.000 0.000 0.000 0.000
## 63 shiai ~~ HR 0.002 0.001 0.001 0.001 0.001
## 64 daritu ~~ steal 0.002 -0.004 -0.004 -0.004 -0.004
## 65 jibun =~ dasuu 0.000 0.000 0.000 0.000 0.000
## 66 jibun =~ daseki 0.000 0.000 0.000 0.000 0.000
## 67 jibun =~ shiai 0.000 0.000 0.000 0.000 0.000
## 68 jibun =~ daritu 0.000 0.000 0.000 0.000 0.000
## 69 jibun =~ steal 0.000 0.000 0.000 0.000 0.000
## 70 jibun =~ datgen 0.000 0.000 0.000 0.000 0.000
## 71 jibun =~ K 0.000 0.000 0.000 0.000 0.000
## 72 aite =~ datgen 0.000 0.000 0.000 0.000 0.000
## 73 aite =~ K 0.000 0.000 0.000 0.000 0.000
## 74 aite =~ HR 0.000 0.000 0.000 0.000 0.000
## 75 aite =~ dead 0.000 0.000 0.000 0.000 0.000
## 76 anda ~~ anda 0.000 0.000 0.000 0.000 0.000
## 77 dasuu ~~ dasuu 0.000 0.000 0.000 0.000 0.000
## 78 daseki ~~ daseki 0.000 0.000 0.000 0.000 0.000
## 79 shiai ~~ shiai 0.000 0.000 0.000 0.000 0.000
## 80 daritu ~~ daritu 0.000 0.000 0.000 0.000 0.000
## 81 steal ~~ steal 0.000 0.000 0.000 0.000 0.000
## 82 datgen ~~ datgen 0.000 0.000 0.000 0.000 0.000
## 83 K ~~ K 0.000 0.000 0.000 0.000 0.000
## 84 four ~~ four 0.000 0.000 0.000 0.000 0.000
## 85 HR ~~ HR 0.000 0.000 0.000 0.000 0.000
## 86 dead ~~ dead 0.000 0.000 0.000 0.000 0.000
## 87 money ~~ money 0.000 0.000 0.000 0.000 0.000
## 88 jibun ~~ jibun 0.000 0.000 0.000 0.000 0.000
## 89 jibun ~~ aite 0.000 0.000 0.000 0.000 0.000
## 90 aite ~~ aite 0.000 0.000 0.000 0.000 0.000
## 91 money ~ jibun 0.000 0.000 0.000 0.000 0.000
## 92 money ~ aite 0.000 0.000 0.000 0.000 0.000
## 93 jibun =~ anda NA NA NA NA NA
## 94 aite =~ four NA NA NA NA NA
## 95 jibun ~ money NA NA NA NA NA
## 96 jibun ~ aite NA NA NA NA NA
## 97 aite ~ money NA NA NA NA NA
## 98 aite ~ jibun NA NA NA NA NA
head(modi[order(modi$mi, decreasing = T), ])
## lhs op rhs mi epc sepc.lv sepc.all sepc.nox
## 1 anda ~~ daritu 36.99 0.125 0.125 0.131 0.131
## 2 daseki ~~ four 26.66 0.048 0.048 0.053 0.053
## 3 datgen ~~ HR 18.01 0.263 0.263 0.263 0.263
## 4 jibun =~ four 16.08 0.612 0.573 0.589 0.589
## 5 jibun =~ HR 13.71 -0.642 -0.602 -0.594 -0.594
## 6 dasuu ~~ four 11.13 -0.030 -0.030 -0.032 -0.032
これでちょっとはやりやすくなったかな？