getwd()
[1] “C:/(中略)/NICER1_3_2/2020-11-24NICER1_3_2/NICER_NNS”
paste(token, collapse="")
myAWL.df <- function(){ # 独自の命令の名前は変えておきましょう
fileV <- NULL
typeV <- NULL
tokenV <- NULL
TTRV <- NULL
GIV <- NULL
NoSV <- NULL
ASLV <- NULL
AWLV <- NULL # AWL用
file.zenbu <- list.files() #
ruiseki <- "" #
for (i in file.zenbu){ #
yomikomi <- readLines(i, warn=F) #
tmp1 <- grep("\\*(JPN|NS)", yomikomi, value=T) #
tmp2 <- gsub("\\*(JPN|NS)...:\t", "", tmp1) #
tmp2b <- gsub("[[:punct:]]", "", tmp2) #
tmp2c <- tolower(tmp2b) #
tmp3 <- strsplit(tmp2c, " ") #
tmp4 <- unlist(tmp3) #
tmp4 <- tmp4[tmp4 != ""] #
token.list <- sort(tmp4) #
type.list <- unique(token.list) #
token <- length(token.list) #
type <- length(type.list) #
TTR <- type/token #
GI <- type/sqrt(token)
NoS <- length(tmp1)
ASL <- token/NoS
mojiretu <- paste(token.list, collapse="") # 長い文字列
mojisuu <- nchar(mojiretu) # 文字数
AWL <- mojisuu/token # 文字数÷単語数
# 各要素の種類ごとにベクトルを作成
fileV <- c(fileV, i)
tokenV <- c(tokenV, token)
typeV <- c(typeV, type)
TTRV <- c(TTRV, TTR)
GIV <- c(GIV, GI)
NoSV <- c(NoSV, NoS)
ASLV <- c(ASLV, ASL)
AWLV <- c(AWLV, AWL) # AWLの追加
}
data.frame(fileV, tokenV, typeV, TTRV, GIV, NoSV, ASLV, AWLV) # 追加修正
}
setwd("NICER_NNS")
NNS.Index.df <- myAWL.df()
names(NNS.Index.df) <- c("ID", "Token", "Type", "TTR", "GI", "NoS", "ASL", "AWL") # 見出しの名前も変えて
head(NNS.Index.df )
## ID Token Type TTR GI NoS ASL AWL
## 1 JPN501.txt 319 134 0.4200627 7.502560 30 10.63333 4.304075
## 2 JPN502.txt 351 158 0.4501425 8.433416 29 12.10345 4.293447
## 3 JPN503.txt 201 121 0.6019900 8.534682 13 15.46154 4.746269
## 4 JPN504.txt 260 139 0.5346154 8.620414 27 9.62963 4.765385
## 5 JPN505.txt 417 174 0.4172662 8.520817 25 16.68000 4.023981
## 6 JPN506.txt 260 123 0.4730769 7.628136 20 13.00000 4.088462
pairs(NNS.Index.df[,-1])
@Topic: education
@EnglishEssay: 4
@SelfEval: 3
@TopicEase: 3
@EssayTraining: 5
@Proctor: 1
@Criterion: 6
@Topic:
の行 grep("@Topic:", 行, value=T)
as.integer()
myIndex.df <- function(){ # 独自の命令の名前は変えておきましょう
topicV <- NULL # topic用
scoreV <- NULL # score用
fileV <- NULL
typeV <- NULL
tokenV <- NULL
TTRV <- NULL
GIV <- NULL
NoSV <- NULL
ASLV <- NULL
AWLV <- NULL
file.zenbu <- list.files() #
ruiseki <- "" #
for (i in file.zenbu){ #
yomikomi <- readLines(i, warn=F) #
topic.tmp <- grep("@Topic:", yomikomi, value=T) # Topicの行
topic <- gsub("@Topic:\t", "", topic.tmp) # 不要部分削除
score.tmp <- grep("@Criterion", yomikomi, value=T) # Scoreの行
score <- gsub("@Criterion:\t", "", score.tmp) # 不要部分削除
tmp1 <- grep("\\*(JPN|NS)", yomikomi, value=T) #
tmp2 <- gsub("\\*(JPN|NS)...:\t", "", tmp1) #
tmp2b <- gsub("[[:punct:]]", "", tmp2) #
tmp2c <- tolower(tmp2b) #
tmp3 <- strsplit(tmp2c, " ") #
tmp4 <- unlist(tmp3) #
tmp4 <- tmp4[tmp4 != ""] #
token.list <- sort(tmp4) #
type.list <- unique(token.list) #
token <- length(token.list) #
type <- length(type.list) #
TTR <- type/token #
GI <- type/sqrt(token)
NoS <- length(tmp1)
ASL <- token/NoS
mojiretu <- paste(token.list, collapse="") #
mojisuu <- nchar(mojiretu) #
AWL <- mojisuu/token #
score <- as.integer(score) # scoreを整数に
# 各要素の種類ごとにベクトルを作成
topicV <- c(topicV, topic) # Topicの追加
scoreV <- c(scoreV, score) # Scoreの追加
fileV <- c(fileV, i)
tokenV <- c(tokenV, token)
typeV <- c(typeV, type)
TTRV <- c(TTRV, TTR)
GIV <- c(GIV, GI)
NoSV <- c(NoSV, NoS)
ASLV <- c(ASLV, ASL)
AWLV <- c(AWLV, AWL)
}
data.frame(fileV, topicV, scoreV, tokenV, typeV, TTRV, GIV, NoSV, ASLV, AWLV) # 追加修正
}
setwd("NICER_NNS")
NNS.Index.df <- myIndex.df()
names(NNS.Index.df) <- c("ID", "Topic", "Score", "Token", "Type", "TTR", "GI", "NoS", "ASL", "AWL") # 見出しの名前も変えて
head(NNS.Index.df)
## ID Topic Score Token Type TTR GI NoS ASL
## 1 JPN501.txt sports 4 319 134 0.4200627 7.502560 30 10.63333
## 2 JPN502.txt education 4 351 158 0.4501425 8.433416 29 12.10345
## 3 JPN503.txt education 3 201 121 0.6019900 8.534682 13 15.46154
## 4 JPN504.txt sports 4 260 139 0.5346154 8.620414 27 9.62963
## 5 JPN505.txt sports 4 417 174 0.4172662 8.520817 25 16.68000
## 6 JPN506.txt money 3 260 123 0.4730769 7.628136 20 13.00000
## AWL
## 1 4.304075
## 2 4.293447
## 3 4.746269
## 4 4.765385
## 5 4.023981
## 6 4.088462
summary(NNS.Index.df)
## ID Topic Score Token
## Length:381 Length:381 Min. :1.000 Min. : 85.0
## Class :character Class :character 1st Qu.:3.000 1st Qu.:209.0
## Mode :character Mode :character Median :3.000 Median :262.0
## Mean :3.522 Mean :275.4
## 3rd Qu.:4.000 3rd Qu.:323.0
## Max. :5.000 Max. :728.0
## NA's :2
## Type TTR GI NoS
## Min. : 49.0 Min. :0.2531 Min. : 4.566 Min. : 7.00
## 1st Qu.:101.0 1st Qu.:0.4230 1st Qu.: 6.947 1st Qu.:17.00
## Median :122.0 Median :0.4699 Median : 7.502 Median :21.00
## Mean :125.6 Mean :0.4697 Mean : 7.582 Mean :22.07
## 3rd Qu.:146.0 3rd Qu.:0.5141 3rd Qu.: 8.279 3rd Qu.:26.00
## Max. :251.0 Max. :0.6581 Max. :10.443 Max. :51.00
##
## ASL AWL
## Min. : 6.96 Min. :3.507
## 1st Qu.:10.82 1st Qu.:4.163
## Median :12.20 Median :4.395
## Mean :12.70 Mean :4.419
## 3rd Qu.:14.08 3rd Qu.:4.652
## Max. :24.00 Max. :5.415
##
str(NNS.Index.df)
## 'data.frame': 381 obs. of 10 variables:
## $ ID : chr "JPN501.txt" "JPN502.txt" "JPN503.txt" "JPN504.txt" ...
## $ Topic: chr "sports" "education" "education" "sports" ...
## $ Score: int 4 4 3 4 4 3 4 3 4 3 ...
## $ Token: int 319 351 201 260 417 260 355 195 260 183 ...
## $ Type : int 134 158 121 139 174 123 149 97 103 99 ...
## $ TTR : num 0.42 0.45 0.602 0.535 0.417 ...
## $ GI : num 7.5 8.43 8.53 8.62 8.52 ...
## $ NoS : int 30 29 13 27 25 20 26 20 19 14 ...
## $ ASL : num 10.63 12.1 15.46 9.63 16.68 ...
## $ AWL : num 4.3 4.29 4.75 4.77 4.02 ...
データフレーム$変数 <- as.factor(データフレーム$変数)
NNS.Index.df$ID <- as.factor(NNS.Index.df$ID)
NNS.Index.df$Topic <- as.factor(NNS.Index.df$Topic)
summary(NNS.Index.df)
## ID Topic Score Token
## JPN501.txt: 1 education:145 Min. :1.000 Min. : 85.0
## JPN502.txt: 1 money : 77 1st Qu.:3.000 1st Qu.:209.0
## JPN503.txt: 1 sports :159 Median :3.000 Median :262.0
## JPN504.txt: 1 Mean :3.522 Mean :275.4
## JPN505.txt: 1 3rd Qu.:4.000 3rd Qu.:323.0
## JPN506.txt: 1 Max. :5.000 Max. :728.0
## (Other) :375 NA's :2
## Type TTR GI NoS
## Min. : 49.0 Min. :0.2531 Min. : 4.566 Min. : 7.00
## 1st Qu.:101.0 1st Qu.:0.4230 1st Qu.: 6.947 1st Qu.:17.00
## Median :122.0 Median :0.4699 Median : 7.502 Median :21.00
## Mean :125.6 Mean :0.4697 Mean : 7.582 Mean :22.07
## 3rd Qu.:146.0 3rd Qu.:0.5141 3rd Qu.: 8.279 3rd Qu.:26.00
## Max. :251.0 Max. :0.6581 Max. :10.443 Max. :51.00
##
## ASL AWL
## Min. : 6.96 Min. :3.507
## 1st Qu.:10.82 1st Qu.:4.163
## Median :12.20 Median :4.395
## Mean :12.70 Mean :4.419
## 3rd Qu.:14.08 3rd Qu.:4.652
## Max. :24.00 Max. :5.415
##
str(NNS.Index.df)
## 'data.frame': 381 obs. of 10 variables:
## $ ID : Factor w/ 381 levels "JPN501.txt","JPN502.txt",..: 1 2 3 4 5 6 7 8 9 10 ...
## $ Topic: Factor w/ 3 levels "education","money",..: 3 1 1 3 3 2 1 3 3 1 ...
## $ Score: int 4 4 3 4 4 3 4 3 4 3 ...
## $ Token: int 319 351 201 260 417 260 355 195 260 183 ...
## $ Type : int 134 158 121 139 174 123 149 97 103 99 ...
## $ TTR : num 0.42 0.45 0.602 0.535 0.417 ...
## $ GI : num 7.5 8.43 8.53 8.62 8.52 ...
## $ NoS : int 30 29 13 27 25 20 26 20 19 14 ...
## $ ASL : num 10.63 12.1 15.46 9.63 16.68 ...
## $ AWL : num 4.3 4.29 4.75 4.77 4.02 ...
hist(NNS.Index.df$Score)
plot(NNS.Index.df$NoS, NNS.Index.df$Token)
cor.test(NNS.Index.df$NoS, NNS.Index.df$Token)
##
## Pearson's product-moment correlation
##
## data: NNS.Index.df$NoS and NNS.Index.df$Token
## t = 23.697, df = 379, p-value < 2.2e-16
## alternative hypothesis: true correlation is not equal to 0
## 95 percent confidence interval:
## 0.7287991 0.8102583
## sample estimates:
## cor
## 0.7726906
setwd("NICER_NS")
NS.Index.df <- myIndex.df()
names(NS.Index.df) <- c("ID", "Topic", "Score", "Token", "Type", "TTR", "GI", "NoS", "ASL", "AWL") # 見出しの名前も変えて
NS.Index.df$ID <- as.factor(NS.Index.df$ID)
NS.Index.df$Topic <- as.factor(NS.Index.df$Topic)
head(NS.Index.df)
## ID Topic Score Token Type TTR GI NoS ASL AWL
## 1 NS501.txt education 5 736 359 0.4877717 13.23292 39 18.87179 4.592391
## 2 NS502.txt education 6 636 340 0.5345912 13.48188 26 24.46154 5.201258
## 3 NS503.txt education 6 834 353 0.4232614 12.22339 22 37.90909 5.565947
## 4 NS504.txt education 6 824 336 0.4077670 11.70511 30 27.46667 5.276699
## 5 NS505.txt sports 6 898 393 0.4376392 13.11458 39 23.02564 4.749443
## 6 NS506.txt education 6 829 339 0.4089264 11.77396 31 26.74194 4.460796
データフレーム内の数値以外の変数は除いていおく
複数ある場合は、コロンで範囲を指定(以下の3通りのいずれか)
pairs(NNS.Index.df[,3:10])
pairs(NS.Index.df[3:10])
始めて使う人は、パソコンにインストールする必要がある。
install.packages(“PerformanceAnalytics”)
一度インストールしたら、あとは、使う前に、library(PerformanceAnalytics) とする。
library(PerformanceAnalytics)
str(NNS.Index.df)
## 'data.frame': 381 obs. of 10 variables:
## $ ID : Factor w/ 381 levels "JPN501.txt","JPN502.txt",..: 1 2 3 4 5 6 7 8 9 10 ...
## $ Topic: Factor w/ 3 levels "education","money",..: 3 1 1 3 3 2 1 3 3 1 ...
## $ Score: int 4 4 3 4 4 3 4 3 4 3 ...
## $ Token: int 319 351 201 260 417 260 355 195 260 183 ...
## $ Type : int 134 158 121 139 174 123 149 97 103 99 ...
## $ TTR : num 0.42 0.45 0.602 0.535 0.417 ...
## $ GI : num 7.5 8.43 8.53 8.62 8.52 ...
## $ NoS : int 30 29 13 27 25 20 26 20 19 14 ...
## $ ASL : num 10.63 12.1 15.46 9.63 16.68 ...
## $ AWL : num 4.3 4.29 4.75 4.77 4.02 ...
chart.Correlation(NNS.Index.df[3:10])
chart.Correlation(NS.Index.df[3:10])
相関関係は、どっちが原因かとか結果とかは考えない。
因果関係は、どっちが原因でどっちが結果かを考える。ただし、判断は分析する人がする。
エッセイのスコアと言語特徴の関係は?
y = ax + b
yが結果
xが原因
散布図の中に直線を引く
plot(NNS.Index.df$Token, NNS.Index.df$Score)
abline(lm(Score ~ Token, data=NNS.Index.df))
lm(結果 ~ 原因, data=データフレーム)
model.1 <- lm(Score ~ Token, data = NNS.Index.df)
summary(model.1)
##
## Call:
## lm(formula = Score ~ Token, data = NNS.Index.df)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.57666 -0.27899 0.00791 0.30202 1.31169
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1.662214 0.070355 23.63 <2e-16 ***
## Token 0.006751 0.000242 27.90 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.4369 on 377 degrees of freedom
## ( 2 個の観測値が欠損のため削除されました )
## Multiple R-squared: 0.6737, Adjusted R-squared: 0.6728
## F-statistic: 778.3 on 1 and 377 DF, p-value: < 2.2e-16
install.packages("effects")
library(effects)
plot(allEffects(model.1))
例えば、時間内にたくさん書けて(Tokenが多い)、難しい単語(AWLが長い)を使う場合
model.2 <- lm(Score ~ Token + AWL, data = NNS.Index.df)
summary(model.2)
##
## Call:
## lm(formula = Score ~ Token + AWL, data = NNS.Index.df)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.45033 -0.25813 0.00101 0.26967 1.18702
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.7612546 0.2840958 -2.680 0.0077 **
## Token 0.0069188 0.0002217 31.214 <2e-16 ***
## AWL 0.5377749 0.0614107 8.757 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.3987 on 376 degrees of freedom
## ( 2 個の観測値が欠損のため削除されました )
## Multiple R-squared: 0.729, Adjusted R-squared: 0.7275
## F-statistic: 505.6 on 2 and 376 DF, p-value: < 2.2e-16
plot(allEffects(model.2))
試しに、全部の要因を入れてみる。
model.3 <- lm(Score ~ Token + Type + TTR + GI + NoS + ASL + AWL, data = NNS.Index.df)
summary(model.3)
##
## Call:
## lm(formula = Score ~ Token + Type + TTR + GI + NoS + ASL + AWL,
## data = NNS.Index.df)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.23548 -0.25526 -0.00781 0.26377 1.48044
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -0.489784 0.625095 -0.784 0.43381
## Token 0.002668 0.002036 1.310 0.19089
## Type -0.010717 0.009695 -1.105 0.26966
## TTR -6.650822 1.236955 -5.377 1.34e-07 ***
## GI 0.625503 0.216334 2.891 0.00406 **
## NoS 0.013683 0.013825 0.990 0.32294
## ASL 0.043424 0.024018 1.808 0.07142 .
## AWL 0.486207 0.060223 8.073 9.64e-15 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.3701 on 371 degrees of freedom
## ( 2 個の観測値が欠損のため削除されました )
## Multiple R-squared: 0.7696, Adjusted R-squared: 0.7653
## F-statistic: 177 on 7 and 371 DF, p-value: < 2.2e-16
VIF値が10以上は適切ではない。
要因間の相関が高い。影響が強く出すぎてしまう。
各要因は「独立」している前提がある。
相関の高い要因のどれかを外す。どれを外すかはよく考える。
相関係数が 0.8以下くらいが目安
https://www.sugiura-ken.org/wiki/wiki.cgi/exp?page=VIF
不要なものを外して、モデルを作り直し、VIFを確認して、10以下にする
パッケージ car に入っている vif() で確認
install.packages("car")
library(car)
vif(model.3)
## Token Type TTR GI NoS ASL AWL
## 98.72870 269.93935 18.13305 123.04640 25.66229 11.97882 1.12481
これではVIFが大きすぎて問題があるので、よく考えよう。(根拠に基づいて合理的に判断)
model.4 <- lm(Score ~ Type + NoS + ASL + AWL, data = NNS.Index.df)
summary(model.4)
##
## Call:
## lm(formula = Score ~ Type + NoS + ASL + AWL, data = NNS.Index.df)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.36867 -0.26243 -0.00216 0.25720 1.22204
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -2.163820 0.320158 -6.759 5.36e-11 ***
## Type 0.005424 0.001347 4.025 6.89e-05 ***
## NoS 0.063868 0.006150 10.385 < 2e-16 ***
## ASL 0.127821 0.011484 11.130 < 2e-16 ***
## AWL 0.445741 0.061970 7.193 3.49e-12 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.3867 on 374 degrees of freedom
## ( 2 個の観測値が欠損のため削除されました )
## Multiple R-squared: 0.7465, Adjusted R-squared: 0.7437
## F-statistic: 275.3 on 4 and 374 DF, p-value: < 2.2e-16
vif(model.4)
## Type NoS ASL AWL
## 4.776499 4.651624 2.508681 1.090959
plot(allEffects(model.4))
Score = 0.005424*Type + 0.063868*NoS + 0.127821*ASL + 0.445741*AWL -2.163820
という関係になる。
どういういうことか、JPN501を参考に考えてみる。
の場合、
Score = 0.005424*134 + 0.063868*30 + 0.127821*10.6 + 0.445741*4.3 -2.163820
を計算するとScoreが予測できる:3.750625
library(easystats)
check_model(model.4)
check_collinearity(model.4)
## # Check for Multicollinearity
##
## Low Correlation
##
## Term VIF VIF 95% CI Increased SE Tolerance Tolerance 95% CI
## Type 4.78 [4.02, 5.72] 2.19 0.21 [0.17, 0.25]
## NoS 4.65 [3.92, 5.57] 2.16 0.21 [0.18, 0.26]
## ASL 2.51 [2.16, 2.96] 1.58 0.40 [0.34, 0.46]
## AWL 1.09 [1.03, 1.33] 1.04 0.92 [0.75, 0.98]
plot(check_collinearity(model.4))
## Variable `Component` is not in your data frame :/
check_normality(model.4)
## OK: residuals appear as normally distributed (p = 0.350).
plot(check_normality(model.4))
report(model.4)
## We fitted a linear model (estimated using OLS) to predict Score with Type, NoS,
## ASL and AWL (formula: Score ~ Type + NoS + ASL + AWL). The model explains a
## statistically significant and substantial proportion of variance (R2 = 0.75,
## F(4, 374) = 275.27, p < .001, adj. R2 = 0.74). The model's intercept,
## corresponding to Type = 0, NoS = 0, ASL = 0 and AWL = 0, is at -2.16 (95% CI
## [-2.79, -1.53], t(374) = -6.76, p < .001). Within this model:
##
## - The effect of Type is statistically significant and positive (beta =
## 5.42e-03, 95% CI [2.77e-03, 8.07e-03], t(374) = 4.03, p < .001; Std. beta =
## 0.23, 95% CI [0.12, 0.34])
## - The effect of NoS is statistically significant and positive (beta = 0.06, 95%
## CI [0.05, 0.08], t(374) = 10.39, p < .001; Std. beta = 0.58, 95% CI [0.47,
## 0.69])
## - The effect of ASL is statistically significant and positive (beta = 0.13, 95%
## CI [0.11, 0.15], t(374) = 11.13, p < .001; Std. beta = 0.46, 95% CI [0.38,
## 0.54])
## - The effect of AWL is statistically significant and positive (beta = 0.45, 95%
## CI [0.32, 0.57], t(374) = 7.19, p < .001; Std. beta = 0.20, 95% CI [0.14,
## 0.25])
##
## Standardized parameters were obtained by fitting the model on a standardized
## version of the dataset. 95% Confidence Intervals (CIs) and p-values were
## computed using a Wald t-distribution approximation.
report_table(model.4)
## Parameter | Coefficient | 95% CI | t(374) | p | Std. Coef. | Std. Coef. 95% CI | Fit
## ------------------------------------------------------------------------------------------------------
## (Intercept) | -2.16 | [-2.79, -1.53] | -6.76 | < .001 | 1.33e-15 | [-0.05, 0.05] |
## Type | 5.42e-03 | [ 0.00, 0.01] | 4.03 | < .001 | 0.23 | [ 0.12, 0.34] |
## NoS | 0.06 | [ 0.05, 0.08] | 10.39 | < .001 | 0.58 | [ 0.47, 0.69] |
## ASL | 0.13 | [ 0.11, 0.15] | 11.13 | < .001 | 0.46 | [ 0.38, 0.54] |
## AWL | 0.45 | [ 0.32, 0.57] | 7.19 | < .001 | 0.20 | [ 0.14, 0.25] |
## | | | | | | |
## AIC | | | | | | | 362.31
## AICc | | | | | | | 362.54
## BIC | | | | | | | 385.94
## R2 | | | | | | | 0.75
## R2 (adj.) | | | | | | | 0.74
## Sigma | | | | | | | 0.39
一人の被験者からは一回だけ(ランダム効果なし)
応答変数の分布の確認をする。
family=分布 という形で、応答変数の分布を指定する
library(tidyverse)
library(openxlsx)
library(lme4)
library(lmerTest)
library(MuMIn)
library(effects)
library(ggplot2)
library(easystats)
model.glm.1 <- glm(Score ~ Type + NoS + ASL + AWL, data = NNS.Index.df, family = poisson)
summary(model.glm.1)
##
## Call:
## glm(formula = Score ~ Type + NoS + ASL + AWL, family = poisson,
## data = NNS.Index.df)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.323186 0.439331 -0.736 0.4620
## Type 0.001548 0.001759 0.880 0.3790
## NoS 0.016669 0.007929 2.102 0.0355 *
## ASL 0.033741 0.015088 2.236 0.0253 *
## AWL 0.130021 0.085635 1.518 0.1289
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for poisson family taken to be 1)
##
## Null deviance: 64.327 on 378 degrees of freedom
## Residual deviance: 19.251 on 374 degrees of freedom
## ( 2 個の観測値が欠損のため削除されました )
## AIC: 1212.2
##
## Number of Fisher Scoring iterations: 4
plot(allEffects(model.glm.1))
model.glm.1.best <- glm(Score ~ NoS + ASL, data = NNS.Index.df, family = poisson)
summary(model.glm.1.best)
##
## Call:
## glm(formula = Score ~ NoS + ASL, family = poisson, data = NNS.Index.df)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.189812 0.174077 1.090 0.276
## NoS 0.021647 0.003790 5.712 1.12e-08 ***
## ASL 0.045374 0.009862 4.601 4.21e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for poisson family taken to be 1)
##
## Null deviance: 64.327 on 378 degrees of freedom
## Residual deviance: 23.001 on 376 degrees of freedom
## ( 2 個の観測値が欠損のため削除されました )
## AIC: 1211.9
##
## Number of Fisher Scoring iterations: 4
plot(allEffects(model.glm.1.best))
summary(NNS.Index.df)
## ID Topic Score Token
## JPN501.txt: 1 education:145 Min. :1.000 Min. : 85.0
## JPN502.txt: 1 money : 77 1st Qu.:3.000 1st Qu.:209.0
## JPN503.txt: 1 sports :159 Median :3.000 Median :262.0
## JPN504.txt: 1 Mean :3.522 Mean :275.4
## JPN505.txt: 1 3rd Qu.:4.000 3rd Qu.:323.0
## JPN506.txt: 1 Max. :5.000 Max. :728.0
## (Other) :375 NA's :2
## Type TTR GI NoS
## Min. : 49.0 Min. :0.2531 Min. : 4.566 Min. : 7.00
## 1st Qu.:101.0 1st Qu.:0.4230 1st Qu.: 6.947 1st Qu.:17.00
## Median :122.0 Median :0.4699 Median : 7.502 Median :21.00
## Mean :125.6 Mean :0.4697 Mean : 7.582 Mean :22.07
## 3rd Qu.:146.0 3rd Qu.:0.5141 3rd Qu.: 8.279 3rd Qu.:26.00
## Max. :251.0 Max. :0.6581 Max. :10.443 Max. :51.00
##
## ASL AWL
## Min. : 6.96 Min. :3.507
## 1st Qu.:10.82 1st Qu.:4.163
## Median :12.20 Median :4.395
## Mean :12.70 Mean :4.419
## 3rd Qu.:14.08 3rd Qu.:4.652
## Max. :24.00 Max. :5.415
##
install.packages("fitdistplus")
library(fitdistrplus)
## 要求されたパッケージ MASS をロード中です
##
## 次のパッケージを付け加えます: 'MASS'
## 以下のオブジェクトは 'package:dplyr' からマスクされています:
##
## select
## 要求されたパッケージ survival をロード中です
summary(NNS.Index.df)
## ID Topic Score Token
## JPN501.txt: 1 education:145 Min. :1.000 Min. : 85.0
## JPN502.txt: 1 money : 77 1st Qu.:3.000 1st Qu.:209.0
## JPN503.txt: 1 sports :159 Median :3.000 Median :262.0
## JPN504.txt: 1 Mean :3.522 Mean :275.4
## JPN505.txt: 1 3rd Qu.:4.000 3rd Qu.:323.0
## JPN506.txt: 1 Max. :5.000 Max. :728.0
## (Other) :375 NA's :2
## Type TTR GI NoS
## Min. : 49.0 Min. :0.2531 Min. : 4.566 Min. : 7.00
## 1st Qu.:101.0 1st Qu.:0.4230 1st Qu.: 6.947 1st Qu.:17.00
## Median :122.0 Median :0.4699 Median : 7.502 Median :21.00
## Mean :125.6 Mean :0.4697 Mean : 7.582 Mean :22.07
## 3rd Qu.:146.0 3rd Qu.:0.5141 3rd Qu.: 8.279 3rd Qu.:26.00
## Max. :251.0 Max. :0.6581 Max. :10.443 Max. :51.00
##
## ASL AWL
## Min. : 6.96 Min. :3.507
## 1st Qu.:10.82 1st Qu.:4.163
## Median :12.20 Median :4.395
## Mean :12.70 Mean :4.419
## 3rd Qu.:14.08 3rd Qu.:4.652
## Max. :24.00 Max. :5.415
##
NNS.Index.df2 <- na.omit(NNS.Index.df)
summary(NNS.Index.df2)
## ID Topic Score Token
## JPN501.txt: 1 education:145 Min. :1.000 Min. : 85.0
## JPN502.txt: 1 money : 77 1st Qu.:3.000 1st Qu.:209.0
## JPN503.txt: 1 sports :157 Median :3.000 Median :262.0
## JPN504.txt: 1 Mean :3.522 Mean :275.6
## JPN505.txt: 1 3rd Qu.:4.000 3rd Qu.:322.5
## JPN506.txt: 1 Max. :5.000 Max. :728.0
## (Other) :373
## Type TTR GI NoS
## Min. : 49.0 Min. :0.2531 Min. : 4.566 Min. : 7.00
## 1st Qu.:101.0 1st Qu.:0.4232 1st Qu.: 6.952 1st Qu.:17.00
## Median :122.0 Median :0.4699 Median : 7.503 Median :21.00
## Mean :125.7 Mean :0.4698 Mean : 7.586 Mean :22.08
## 3rd Qu.:146.0 3rd Qu.:0.5137 3rd Qu.: 8.283 3rd Qu.:26.00
## Max. :251.0 Max. :0.6581 Max. :10.443 Max. :51.00
##
## ASL AWL
## Min. : 6.96 Min. :3.507
## 1st Qu.:10.81 1st Qu.:4.163
## Median :12.21 Median :4.395
## Mean :12.71 Mean :4.420
## 3rd Qu.:14.11 3rd Qu.:4.652
## Max. :24.00 Max. :5.415
##
descdist(NNS.Index.df2$Score, boot=500)
## summary statistics
## ------
## min: 1 max: 5
## median: 3
## mean: 3.522427
## estimated sd: 0.7638654
## estimated skewness: 0.01361058
## estimated kurtosis: 3.021945
model.glm.1 <- glm(Score ~ Type + NoS + ASL + AWL, data = NNS.Index.df2, family = gaussian)
summary(model.glm.1)
##
## Call:
## glm(formula = Score ~ Type + NoS + ASL + AWL, family = gaussian,
## data = NNS.Index.df2)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -2.163820 0.320158 -6.759 5.36e-11 ***
## Type 0.005424 0.001347 4.025 6.89e-05 ***
## NoS 0.063868 0.006150 10.385 < 2e-16 ***
## ASL 0.127821 0.011484 11.130 < 2e-16 ***
## AWL 0.445741 0.061970 7.193 3.49e-12 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.1495256)
##
## Null deviance: 220.559 on 378 degrees of freedom
## Residual deviance: 55.923 on 374 degrees of freedom
## AIC: 362.31
##
## Number of Fisher Scoring iterations: 2
model.glm.2 <- glm(Score ~ Type * NoS * ASL * AWL, data = NNS.Index.df2, family = gaussian)
summary(model.glm.2)
##
## Call:
## glm(formula = Score ~ Type * NoS * ASL * AWL, family = gaussian,
## data = NNS.Index.df2)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -2.519e+01 1.403e+01 -1.796 0.0733 .
## Type 1.536e-01 1.204e-01 1.276 0.2026
## NoS 1.393e+00 6.831e-01 2.040 0.0421 *
## ASL 1.857e+00 1.019e+00 1.823 0.0691 .
## AWL 5.377e+00 3.244e+00 1.657 0.0983 .
## Type:NoS -7.408e-03 4.973e-03 -1.490 0.1371
## Type:ASL -9.109e-03 8.453e-03 -1.078 0.2819
## NoS:ASL -1.127e-01 5.283e-02 -2.133 0.0336 *
## Type:AWL -2.989e-02 2.767e-02 -1.080 0.2807
## NoS:AWL -2.969e-01 1.570e-01 -1.891 0.0595 .
## ASL:AWL -3.863e-01 2.355e-01 -1.641 0.1017
## Type:NoS:ASL 5.509e-04 3.650e-04 1.510 0.1320
## Type:NoS:AWL 1.567e-03 1.143e-03 1.371 0.1712
## Type:ASL:AWL 1.889e-03 1.946e-03 0.971 0.3323
## NoS:ASL:AWL 2.614e-02 1.211e-02 2.158 0.0316 *
## Type:NoS:ASL:AWL -1.221e-04 8.400e-05 -1.453 0.1470
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.1369537)
##
## Null deviance: 220.559 on 378 degrees of freedom
## Residual deviance: 49.714 on 363 degrees of freedom
## AIC: 339.71
##
## Number of Fisher Scoring iterations: 2
library(MuMIn)
options(na.action = "na.fail")
dredge(model.glm.2, rank="AIC")
## Fixed term is "(Intercept)"
## Global model call: glm(formula = Score ~ Type * NoS * ASL * AWL, family = gaussian,
## data = NNS.Index.df2)
## ---
## Model selection table
## (Int) ASL AWL NoS Typ ASL:AWL ASL:NoS
## 624 -3.182000 0.11930 0.464700 0.0619000 0.0196000 0.004805
## 752 -2.158000 0.12410 0.230100 0.0141800 0.0189200 0.004523
## 1792 -8.813000 0.63180 1.765000 0.3610000 0.0183700 -0.1157000 -0.022530
## 880 -2.262000 0.12160 0.258100 0.0599100 0.0119400 0.004868
## 2048 -8.944000 0.60020 1.798000 0.3423000 0.0355900 -0.1089000 -0.026990
## 4720 -3.659000 0.15550 0.464400 0.0865900 0.0231900 0.002877
## 640 -3.484000 0.14280 0.531700 0.0630400 0.0194400 -0.0051990 0.004712
## 5888 -9.252000 0.67280 1.716000 0.3890000 0.0232700 -0.1137000 -0.025100
## 4848 -2.725000 0.17330 0.212200 0.0438900 0.0237100 0.001904
## 768 -1.920000 0.10880 0.177100 0.0115900 0.0190000 0.0034120 0.004573
## 1008 -2.139000 0.12410 0.226100 0.0154000 0.0185500 0.004534
## 4976 -2.776000 0.16650 0.241000 0.0901800 0.0157500 0.002497
## 896 -2.799000 0.17580 0.376800 0.0621800 0.0103200 -0.0118800 0.004666
## 4096 -6.916000 0.44180 1.330000 0.4551000 0.0005696 -0.0724600 -0.035980
## 6144 -9.443000 0.64350 1.746000 0.3721000 0.0423700 -0.1061000 -0.030200
## 10240 -8.481000 0.60280 1.688000 0.3207000 0.0323000 -0.1093000 -0.027460
## 4736 -3.911000 0.17520 0.522800 0.0869800 0.0229600 -0.0045310 0.002843
## 1728 -9.121000 0.65940 1.955000 0.3938000 0.0106000 -0.1301000 -0.025370
## 560 -2.582000 0.07901 0.462500 0.0807200 0.0114900 0.002929
## 4864 -2.385000 0.15190 0.131300 0.0410300 0.0239900 0.0051710 0.001887
## 5104 -2.695000 0.17340 0.205400 0.0460200 0.0230900 0.001916
## 688 -1.475000 0.08547 0.204900 0.0277500 0.0109900 0.002674
## 8192 -7.267000 0.47400 1.225000 0.4994000 0.0045930 -0.0657600 -0.040430
## 9200 -2.695000 0.12320 0.358300 0.0419200 0.0229900 0.004535
## 1024 -1.781000 0.09227 0.145400 0.0008868 0.0214800 0.0071270 0.004557
## 4992 -3.283000 0.21780 0.356000 0.0916800 0.0140800 -0.0114800 0.002356
## 1984 -9.270000 0.61850 1.988000 0.3688000 0.0324600 -0.1209000 -0.030810
## 2944 -5.108000 0.34880 0.909100 0.0608300 0.0299500 -0.0515200 0.004738
## 12288 -5.971000 0.43570 1.106000 0.4249000 -0.0075300 -0.0707800 -0.037410
## 14336 -9.144000 0.64450 1.677000 0.3580000 0.0402000 -0.1064000 -0.030450
## 576 -3.520000 0.15170 0.662400 0.0829800 0.0115200 -0.0155100 0.002765
## 816 -1.906000 0.07954 0.314500 0.0798400 0.0057820 0.002920
## 944 -1.754000 0.08716 0.260700 0.0105500 0.0163600 0.002596
## 13296 -3.496000 0.17460 0.388300 0.0844200 0.0295000 0.001787
## 5120 -2.192000 0.12880 0.084060 0.0261200 0.0277100 0.0106200 0.001803
## 704 -1.992000 0.11850 0.316700 0.0328200 0.0110400 -0.0071610 0.002618
## 3072 -4.432000 0.28990 0.757200 -0.0100600 0.0470400 -0.0381600 0.004628
## 32768 -25.190000 1.85700 5.377000 1.3930000 0.1536000 -0.3863000 -0.112700
## 16384 -6.500000 0.46830 1.048000 0.4741000 -0.0020090 -0.0645600 -0.041470
## 7040 -5.362000 0.37330 0.844200 0.0884800 0.0317900 -0.0477500 0.002576
## 9216 -2.320000 0.09700 0.273000 0.0261800 0.0247900 0.0058920 0.004554
## 832 -2.911000 0.18130 0.526900 0.0825900 0.0032420 -0.0216600 0.002687
## 10176 -9.418000 0.61760 2.023000 0.3757000 0.0335400 -0.1207000 -0.030650
## 7168 -4.655000 0.31210 0.660500 0.0145400 0.0513400 -0.0320000 0.002013
## 9136 -2.966000 0.08600 0.546400 0.0680900 0.0260400 0.002641
## 13312 -2.953000 0.13660 0.258800 0.0619600 0.0325000 0.0090190 0.001710
## 960 -1.649000 0.07787 0.237300 0.0063300 0.0172100 0.0020700 0.002600
## 11264 -5.047000 0.29670 0.902300 0.0176400 0.0508800 -0.0398900 0.004625
## 656 -1.618000 0.13530 0.107900 0.0354600 0.0112800
## 528 -3.132000 0.13300 0.443400 0.1080000 0.0120000
## 912 -2.120000 0.13570 0.215200 0.0035790 0.0210900
## 15360 -5.479000 0.32190 0.849800 0.0521900 0.0566000 -0.0340700 0.001917
## 672 -2.752000 0.20590 0.359200 0.0462700 0.0113800 -0.0158000
## 9152 -3.027000 0.09027 0.560000 0.0706100 0.0257500 -0.0009527 0.002640
## 544 -4.694000 0.25300 0.789700 0.1092000 0.0119900 -0.0267300
## 720 -1.743000 0.14690 0.101200 0.0341400 0.0125800
## 800 -3.921000 0.28520 0.622300 0.1079000 0.0020130 -0.0337400
## 784 -2.420000 0.13330 0.288000 0.1070000 0.0059970
## 592 -3.245000 0.14190 0.442200 0.1078000 0.0130000
## 976 -2.212000 0.14590 0.205600 0.0035290 0.0218900
## 9104 -2.786000 0.13550 0.370800 0.0349400 0.0264200
## 928 -2.275000 0.14940 0.250000 0.0098660 0.0198200 -0.0030740
## 736 -2.823000 0.21390 0.343800 0.0446700 0.0125500 -0.0152100
## 608 -4.777000 0.25950 0.785400 0.1091000 0.0128600 -0.0264800
## 864 -4.006000 0.29470 0.611500 0.1076000 0.0028450 -0.0336300
## 848 -2.515000 0.14400 0.277900 0.1067000 0.0068570
## 9168 -2.733000 0.14510 0.329500 0.0283700 0.0260600
## 992 -2.359000 0.15880 0.238600 0.0094800 0.0206900 -0.0029090
## 9120 -3.098000 0.15740 0.441600 0.0480300 0.0249100 -0.0049140
## 2912 -4.950000 0.36560 0.828800 0.1074000 0.0107600 -0.0498300
## 3040 -3.974000 0.27950 0.611000 0.0029130 0.0362000 -0.0305100
## 9184 -3.013000 0.16450 0.393400 0.0401800 0.0247100 -0.0044010
## 11232 -4.688000 0.28740 0.779700 0.0350900 0.0406800 -0.0325100
## 1280 -9.217000 0.80090 2.071000 0.4228000 0.0093440 -0.1539000 -0.034260
## 1216 -9.410000 0.81100 2.188000 0.4423000 0.0044300 -0.1619000 -0.035610
## 1536 -9.343000 0.77100 2.103000 0.4051000 0.0257200 -0.1475000 -0.038530
## 1472 -9.546000 0.77510 2.218000 0.4201000 0.0240800 -0.1538000 -0.040560
## 3584 -7.514000 0.62850 1.682000 0.5071000 -0.0059070 -0.1147000 -0.046670
## 240 -0.255800 0.12610 0.023360 -0.0572800 0.0090730 0.002749
## 176 0.090520 0.10090 0.017580 -0.0447600 0.0044180 0.001637
## 144 -0.105000 0.13160 -0.031820 -0.0350000 0.0050420
## 368 -0.137400 0.12230 0.007903 0.0269600 -0.0065260 0.003397
## 496 0.004104 0.12590 -0.034920 -0.0384200 0.0036140 0.002929
## 256 0.482000 0.07798 -0.142300 -0.0650100 0.0093730 0.0107500 0.002918
## 192 0.315800 0.08611 -0.031590 -0.0468000 0.0044150 0.0031820 0.001665
## 432 0.170000 0.10030 0.001223 -0.0392500 0.0028030 0.001665
## 208 -0.200300 0.14080 -0.037340 -0.0361700 0.0060520
## 304 0.019650 0.09136 0.061530 0.0432100 -0.0101500 0.002039
## 400 -0.184900 0.13170 -0.013760 -0.0409600 0.0067270
## 160 -0.322700 0.14560 0.017270 -0.0331600 0.0050350 -0.0031230
## 384 -1.053000 0.21220 0.207700 0.0311100 -0.0090180 -0.0197100 0.003078
## 320 -1.215000 0.21430 0.321100 0.0469100 -0.0130600 -0.0262000 0.001766
## 112 -2.061000 0.11690 0.460100 0.0286000 0.0094030 0.003121
## 2432 -5.394000 0.53430 1.204000 0.0291900 0.0279900 -0.0935200 0.003242
## 48 -1.725000 0.09093 0.458900 0.0423700 0.0046300 0.001984
## 288 -2.003000 0.28270 0.398200 0.0662300 -0.0129000 -0.0341200
## 512 0.274600 0.10200 -0.095720 -0.0493600 0.0058140 0.0053640 0.002946
## 272 -0.482700 0.12920 0.059930 0.0652600 -0.0088900
## 448 0.268600 0.09154 -0.020810 -0.0432300 0.0036010 0.0019500 0.001668
## 464 -0.263700 0.14060 -0.022360 -0.0410400 0.0074130
## 224 -0.380000 0.15220 0.003764 -0.0346100 0.0060250 -0.0026070
## 416 -0.257800 0.13810 0.002659 -0.0379900 0.0061270 -0.0014520
## 2560 -4.465000 0.45130 0.992500 -0.0675500 0.0514100 -0.0747300 0.003110
## 16 -2.164000 0.12780 0.445700 0.0638700 0.0054240
## 1208 -9.227000 0.80860 2.165000 0.4357000 -0.1599000 -0.034390
## 128 -2.394000 0.14280 0.533700 0.0298600 0.0092320 -0.0057130 0.003018
## 64 -2.464000 0.14850 0.617000 0.0440300 0.0046270 -0.0122700 0.001851
## 352 -2.076000 0.29110 0.388600 0.0659800 -0.0121900 -0.0340100
## 336 -0.564500 0.13860 0.050830 0.0650000 -0.0081480
## 32 -3.329000 0.21830 0.707000 0.0642200 0.0053320 -0.0201600
## 480 -0.329200 0.14640 -0.007571 -0.0383800 0.0068720 -0.0013060
## 80 -2.224000 0.13270 0.445100 0.0637200 0.0059630
## 2400 -5.251000 0.52680 1.115000 0.0659400 0.0144200 -0.0878700
## 2528 -4.139000 0.42720 0.864800 -0.0524000 0.0433100 -0.0655200
## 96 -3.370000 0.22160 0.704700 0.0641000 0.0057620 -0.0200200
## 168 0.141900 0.10750 0.023920 -0.0481900 0.002671
## 184 0.452300 0.08714 -0.043860 -0.0510000 0.0043860 0.002709
## 40 -1.842000 0.09696 0.506900 0.0469200 0.003104
## 136 -0.195300 0.16420 -0.062960 -0.0315900
## 56 -2.596000 0.15570 0.668400 0.0486100 -0.0125300 0.002968
## 152 -0.652100 0.19340 0.040200 -0.0277400 -0.0065570
## 8 -2.635000 0.16270 0.498400 0.0855100
## 24 -4.159000 0.28100 0.841100 0.0854900 -0.0265300
## 622 -1.230000 0.12720 0.0610000 0.0215500 0.003734
## 4718 -1.782000 0.16910 0.0895100 0.0257000 0.001509
## 558 -0.668900 0.08893 0.0788900 0.0138500 0.001957
## 526 -1.095000 0.12520 0.0974200 0.0141200
## 590 -1.354000 0.14470 0.0971300 0.0163000
## 14 -0.102700 0.11990 0.0526400 0.0074700
## 46 0.157900 0.10060 0.0412400 0.0070870 0.001036
## 110 -0.150800 0.12480 0.0283900 0.0115500 0.002095
## 78 -0.303700 0.13540 0.0522200 0.0091660
## 38 0.289200 0.11200 0.0484100 0.002691
## 6 -0.431500 0.16880 0.0819200
## 12 -0.452700 0.03999 0.282400 0.0176600
## 76 -0.979900 0.08136 0.280200 0.0218200
## 268 -0.121300 0.03990 0.207300 0.0149500
## 28 -0.795300 0.06666 0.359500 0.0176500 -0.0059730
## 332 -0.489300 0.08277 0.164500 0.0178100
## 92 -1.257000 0.10290 0.343200 0.0218000 -0.0048760
## 284 -0.508900 0.07882 0.294400 0.0140100 -0.0087270
## 348 -0.875200 0.12150 0.251200 0.0168800 -0.0086900
## 2396 -4.488000 0.38980 1.078000 0.0471300 -0.0699800
## 527 -1.233000 0.376300 0.0450000 0.0219900
## 655 -0.874200 0.298200 0.0279300 0.0218600
## 783 -1.131000 0.354200 0.0448300 0.0211400
## 911 -1.045000 0.335100 0.0170200 0.0252300
## 9103 -3.635000 0.938700 0.1387000 0.0459100
## 15 -0.556200 0.380000 0.0134700 0.0167200
## 143 0.306400 0.183700 -0.0276300 0.0167000
## 271 0.342900 0.175400 0.0139100 0.0092150
## 74 0.164400 0.08942 0.0220500
## 10 0.739200 0.04512 0.0175900
## 399 0.436000 0.154200 -0.0179500 0.0139600
## 11 -0.276800 0.325600 0.0187800
## 267 0.245100 0.207000 0.0145100
## 525 0.416500 0.0390900 0.0233100
## 9 1.151000 0.0188700
## 13 1.132000 0.0064760 0.0178900
## 7 -0.795400 0.627900 0.0698500
## 135 0.396700 0.356300 0.0129900
## 5 2.098000 0.0645400
## 4 1.040000 0.11010 0.245200
## 20 -0.662800 0.24220 0.627900 -0.0296300
## 2 2.070000 0.11430
## 3 1.879000 0.371700
## 1 3.522000
## ASL:Typ AWL:NoS AWL:Typ NoS:Typ ASL:AWL:NoS ASL:AWL:Typ
## 624 -5.933e-04 -3.595e-04
## 752 -5.776e-04 0.011230 -3.418e-04
## 1792 -5.701e-04 -0.068730 -3.198e-04 0.006166
## 880 -6.173e-04 1.711e-03 -3.460e-04
## 2048 -5.488e-04 -0.064300 -3.930e-03 -3.203e-04 0.007155
## 4720 -8.582e-04 -5.340e-04
## 640 -5.812e-04 -3.594e-04
## 5888 -9.352e-04 -0.067470 -5.571e-04 0.006151
## 4848 -9.334e-04 0.012060 -5.757e-04
## 768 -5.849e-04 0.011660 -3.412e-04
## 1008 -5.792e-04 0.010910 8.766e-05 -3.416e-04
## 4976 -9.457e-04 1.850e-03 -5.601e-04
## 896 -5.936e-04 1.995e-03 -3.437e-04
## 4096 2.150e-03 -0.089660 4.031e-03 -3.212e-04 0.009170 -0.0006108
## 6144 -9.543e-04 -0.062550 -4.232e-03 -5.850e-04 0.007214
## 10240 -5.541e-04 -0.059230 -3.145e-03 -1.388e-04 0.007262
## 4736 -8.411e-04 -5.296e-04
## 1728 -0.071770 -2.953e-04 0.006379
## 560 -3.333e-04
## 4864 -9.571e-04 0.012750 -5.831e-04
## 5104 -9.370e-04 0.011540 1.467e-04 -5.760e-04
## 688 0.012330 -3.147e-04
## 8192 1.970e-03 -0.090250 4.491e-03 -6.144e-04 0.009436 -0.0006718
## 9200 -5.734e-04 0.004685 -9.640e-04 -5.413e-04
## 1024 -5.822e-04 0.014140 -5.626e-04 -3.418e-04
## 4992 -9.153e-04 2.121e-03 -5.529e-04
## 1984 -0.066090 -4.908e-03 -2.971e-04 0.007604
## 2944 -2.046e-03 -2.491e-03 -3.401e-04 0.0003295
## 12288 2.321e-03 -0.082500 5.937e-03 -3.851e-06 0.009492 -0.0006518
## 14336 -9.520e-04 -0.059370 -3.731e-03 -4.665e-04 0.007281
## 576 -3.347e-04
## 816 1.225e-03 -3.230e-04
## 944 0.016540 -1.189e-03 -3.185e-04
## 13296 -9.467e-04 0.002911 -1.313e-03 -8.654e-04
## 5120 -9.615e-04 0.016380 -8.186e-04 -5.894e-04
## 704 0.011400 -3.168e-04
## 3072 -2.345e-03 0.016290 -6.403e-03 -3.372e-04 0.0004005
## 32768 -9.109e-03 -0.296900 -2.989e-02 -7.408e-03 0.026140 0.0018890
## 16384 2.112e-03 -0.084490 6.012e-03 -3.526e-04 0.009689 -0.0007033
## 7040 -2.222e-03 -1.990e-03 -5.357e-04 0.0003013
## 9216 -5.766e-04 0.008231 -1.353e-03 -5.132e-04
## 832 1.777e-03 -3.202e-04
## 10176 -0.067720 -5.159e-03 -3.558e-04 0.007568
## 7168 -2.594e-03 0.018280 -6.278e-03 -5.722e-04 0.0003752
## 9136 0.002920 -3.438e-03 -7.515e-04
## 13312 -9.661e-04 0.008271 -1.922e-03 -8.349e-04
## 960 0.017480 -1.380e-03 -3.185e-04
## 11264 -2.353e-03 0.009811 -7.319e-03 -5.256e-04 0.0004037
## 656 0.016160 -2.818e-04
## 528 -3.027e-04
## 912 0.023650 -2.175e-03 -2.904e-04
## 15360 -2.613e-03 0.009751 -7.489e-03 -8.307e-04 0.0003786
## 672 0.013920 -2.879e-04
## 9152 0.002349 -3.373e-03 -7.558e-04
## 544 -3.081e-04
## 720 -9.240e-05 0.016410 -2.825e-04
## 800 2.139e-03 -2.915e-04
## 784 1.287e-03 -2.919e-04
## 592 -7.061e-05 -3.035e-04
## 976 -8.102e-05 0.023610 -2.099e-03 -2.907e-04
## 9104 0.016280 -3.413e-03 -5.267e-04
## 928 0.022230 -1.890e-03 -2.905e-04
## 736 -8.382e-05 0.014230 -2.883e-04
## 608 -6.072e-05 -3.087e-04
## 864 -8.003e-05 2.205e-03 -2.918e-04
## 848 -8.416e-05 1.359e-03 -2.922e-04
## 9168 -7.541e-05 0.017770 -3.085e-03 -4.778e-04
## 992 -8.084e-05 0.022260 -1.830e-03 -2.908e-04
## 9120 0.013300 -3.077e-03 -5.497e-04
## 2912 -6.655e-04 3.871e-04 -2.901e-04 0.0001336
## 3040 -1.147e-03 0.023640 -5.388e-03 -2.875e-04 0.0002432
## 9184 -7.451e-05 0.015080 -2.788e-03 -4.990e-04
## 11232 -1.157e-03 0.016110 -6.453e-03 -5.065e-04 0.0002470
## 1280 -3.859e-04 -0.089400 0.008491
## 1216 -0.090400 0.008516
## 1536 -3.652e-04 -0.085210 -3.742e-03 0.009436
## 1472 -0.085390 -4.419e-03 0.009630
## 3584 2.071e-03 -0.108200 3.443e-03 0.011260 -0.0005513
## 240 -3.649e-04 0.020920
## 176 0.021140
## 144 0.023000
## 368 -4.447e-04 3.748e-03
## 496 -3.915e-04 0.016190 1.302e-03
## 256 -3.893e-04 0.022230
## 192 0.021530
## 432 0.019820 3.628e-04
## 208 -7.259e-05 0.023200
## 304 3.291e-03
## 400 0.024340 -3.809e-04
## 160 0.022580
## 384 -4.074e-04 4.196e-03
## 320 3.938e-03
## 112 -3.743e-04
## 2432 -3.120e-03 -4.215e-03 0.0006147
## 48
## 288 4.052e-03
## 512 -3.937e-04 0.018620 8.129e-04
## 272 3.193e-03
## 448 0.020720 1.831e-04
## 464 -7.080e-05 0.024310 -3.132e-04
## 224 -7.104e-05 0.022850
## 416 0.023670 -2.460e-04
## 2560 -3.517e-03 0.022310 -9.553e-03 0.0007086
## 16
## 1208 -0.087840 0.008475
## 128 -3.611e-04
## 64
## 352 -6.982e-05 4.111e-03
## 336 -7.397e-05 3.258e-03
## 32
## 480 -7.072e-05 0.023700 -1.919e-04
## 80 -3.866e-05
## 2400 -2.016e-03 -1.968e-03 0.0004440
## 2528 -2.549e-03 0.026880 -8.509e-03 0.0005655
## 96 -3.076e-05
## 168 0.023020
## 184 0.023560
## 40
## 136 0.026820
## 56
## 152 0.025940
## 8
## 24
## 622 -5.636e-04 -3.521e-04
## 4718 -8.695e-04 -5.536e-04
## 558 -3.273e-04
## 526 -3.068e-04
## 590 -1.541e-04 -3.084e-04
## 14
## 46
## 110 -3.493e-04
## 78 -1.221e-04
## 38
## 6
## 12
## 76 -3.153e-04
## 268 6.158e-04
## 28
## 332 -3.272e-04 9.485e-04
## 92 -3.137e-04
## 284 8.258e-04
## 348 -3.271e-04 1.158e-03
## 2396 -2.541e-03 -5.755e-03 0.0005051
## 527 -2.268e-04
## 655 0.003747 -2.216e-04
## 783 1.825e-04 -2.252e-04
## 911 0.006295 -7.441e-04 -2.245e-04
## 9103 -0.022230 -5.558e-03 -1.142e-03
## 15
## 143 0.009417
## 271 1.690e-03
## 74 -3.379e-04
## 10
## 399 0.007236 6.191e-04
## 11
## 267 9.711e-04
## 525 -2.340e-04
## 9
## 13
## 7
## 135 0.013010
## 5
## 4
## 20
## 2
## 3
## 1
## ASL:NoS:Typ AWL:NoS:Typ ASL:AWL:NoS:Typ df logLik AIC delta weight
## 624 9 -157.854 333.7 0.00 0.106
## 752 10 -157.073 334.1 0.44 0.085
## 1792 12 -155.074 334.1 0.44 0.085
## 880 10 -157.419 334.8 1.13 0.060
## 2048 13 -154.691 335.4 1.67 0.046
## 4720 1.329e-05 10 -157.721 335.4 1.73 0.045
## 640 10 -157.825 335.7 1.94 0.040
## 5888 1.819e-05 13 -154.828 335.7 1.95 0.040
## 4848 1.791e-05 11 -156.834 335.7 1.96 0.040
## 768 11 -157.061 336.1 2.41 0.032
## 1008 11 -157.072 336.1 2.44 0.031
## 4976 1.638e-05 11 -157.218 336.4 2.73 0.027
## 896 11 -157.280 336.6 2.85 0.026
## 4096 14 -154.361 336.7 3.01 0.024
## 6144 2.029e-05 14 -154.387 336.8 3.07 0.023
## 10240 -4.245e-05 14 -154.673 337.3 3.64 0.017
## 4736 1.296e-05 11 -157.699 337.4 3.69 0.017
## 1728 11 -157.723 337.4 3.74 0.016
## 560 8 -160.804 337.6 3.90 0.015
## 4864 1.855e-05 12 -156.808 337.6 3.91 0.015
## 5104 1.796e-05 12 -156.832 337.7 3.96 0.015
## 688 9 -159.873 337.7 4.04 0.014
## 8192 2.247e-05 15 -153.990 338.0 4.27 0.013
## 9200 4.679e-05 12 -157.049 338.1 4.39 0.012
## 1024 12 -157.053 338.1 4.40 0.012
## 4992 1.601e-05 12 -157.089 338.2 4.47 0.011
## 1984 12 -157.128 338.3 4.55 0.011
## 2944 12 -157.128 338.3 4.55 0.011
## 12288 -7.425e-05 15 -154.308 338.6 4.91 0.009
## 14336 2.001e-05 -2.685e-05 15 -154.379 338.8 5.05 0.009
## 576 9 -160.541 339.1 5.37 0.007
## 816 9 -160.582 339.2 5.46 0.007
## 944 10 -159.771 339.5 5.83 0.006
## 13296 1.885e-05 6.507e-05 13 -156.788 339.6 5.87 0.006
## 5120 1.897e-05 13 -156.790 339.6 5.87 0.006
## 704 10 -159.822 339.6 5.94 0.005
## 3072 13 -156.833 339.7 5.96 0.005
## 32768 5.509e-04 1.567e-03 -0.0001221 17 -152.857 339.7 6.01 0.005
## 16384 2.194e-05 -5.965e-05 16 -153.956 339.9 6.20 0.005
## 7040 1.494e-05 13 -156.963 339.9 6.22 0.005
## 9216 4.019e-05 13 -157.036 340.1 6.36 0.004
## 832 10 -160.114 340.2 6.52 0.004
## 10176 1.371e-05 13 -157.126 340.3 6.54 0.004
## 7168 1.798e-05 14 -156.597 341.2 7.49 0.003
## 9136 1.013e-04 11 -159.663 341.3 7.62 0.002
## 13312 1.958e-05 5.566e-05 14 -156.758 341.5 7.81 0.002
## 960 11 -159.769 341.5 7.83 0.002
## 11264 4.417e-05 14 -156.813 341.6 7.92 0.002
## 656 8 -162.889 341.8 8.07 0.002
## 528 7 -164.508 343.0 9.31 0.001
## 912 9 -162.546 343.1 9.38 0.001
## 15360 1.861e-05 5.863e-05 15 -156.562 343.1 9.42 0.001
## 672 9 -162.639 343.3 9.57 0.001
## 9152 1.024e-04 12 -159.663 343.3 9.62 0.001
## 544 8 -163.709 343.4 9.71 0.001
## 720 9 -162.783 343.6 9.86 0.001
## 800 9 -163.096 344.2 10.48 0.001
## 784 8 -164.268 344.5 10.83 0.000
## 592 8 -164.446 344.9 11.19 0.000
## 976 10 -162.464 344.9 11.22 0.000
## 9104 5.536e-05 10 -162.514 345.0 11.32 0.000
## 928 10 -162.542 345.1 11.38 0.000
## 736 10 -162.552 345.1 11.40 0.000
## 608 9 -163.663 345.3 11.62 0.000
## 864 10 -163.017 346.0 12.33 0.000
## 848 9 -164.181 346.4 12.65 0.000
## 9168 4.385e-05 11 -162.444 346.9 13.18 0.000
## 992 11 -162.461 346.9 13.21 0.000
## 9120 6.073e-05 11 -162.505 347.0 13.30 0.000
## 2912 11 -162.992 348.0 14.28 0.000
## 3040 12 -162.381 348.8 15.05 0.000
## 9184 4.879e-05 12 -162.437 348.9 15.17 0.000
## 11232 5.131e-05 13 -162.355 350.7 17.00 0.000
## 1280 11 -166.006 354.0 20.30 0.000
## 1216 10 -167.185 354.4 20.66 0.000
## 1536 12 -165.679 355.4 21.65 0.000
## 1472 11 -166.726 355.5 21.74 0.000
## 3584 13 -165.425 356.8 23.14 0.000
## 240 9 -169.745 357.5 23.78 0.000
## 176 8 -170.827 357.7 23.95 0.000
## 144 7 -171.945 357.9 24.18 0.000
## 368 9 -170.350 358.7 24.99 0.000
## 496 10 -169.633 359.3 25.56 0.000
## 256 10 -169.640 359.3 25.57 0.000
## 192 9 -170.817 359.6 25.93 0.000
## 432 9 -170.818 359.6 25.93 0.000
## 208 8 -171.882 359.8 26.06 0.000
## 304 8 -171.919 359.8 26.13 0.000
## 400 8 -171.935 359.9 26.16 0.000
## 160 8 -171.935 359.9 26.16 0.000
## 384 10 -169.993 360.0 26.28 0.000
## 320 9 -171.274 360.5 26.84 0.000
## 112 8 -172.387 360.8 27.07 0.000
## 2432 11 -169.492 361.0 27.28 0.000
## 48 7 -173.510 361.0 27.31 0.000
## 288 8 -172.532 361.1 27.36 0.000
## 512 11 -169.623 361.2 27.54 0.000
## 272 7 -173.672 361.3 27.64 0.000
## 448 10 -170.816 361.6 27.93 0.000
## 464 9 -171.875 361.8 28.04 0.000
## 224 9 -171.876 361.8 28.04 0.000
## 416 9 -171.934 361.9 28.16 0.000
## 2560 12 -168.971 361.9 28.23 0.000
## 16 6 -175.157 362.3 28.61 0.000
## 1208 9 -172.253 362.5 28.80 0.000
## 128 9 -172.355 362.7 29.00 0.000
## 64 8 -173.356 362.7 29.00 0.000
## 352 9 -172.475 362.9 29.24 0.000
## 336 8 -173.608 363.2 29.51 0.000
## 32 7 -174.725 363.5 29.74 0.000
## 480 10 -171.875 363.7 30.04 0.000
## 80 7 -175.139 364.3 30.57 0.000
## 2400 10 -172.213 364.4 30.72 0.000
## 2528 11 -171.460 364.9 31.21 0.000
## 96 8 -174.714 365.4 31.72 0.000
## 168 7 -175.774 365.5 31.84 0.000
## 184 8 -175.757 367.5 33.81 0.000
## 40 6 -178.884 369.8 36.06 0.000
## 136 6 -178.940 369.9 36.17 0.000
## 56 7 -178.728 371.5 37.75 0.000
## 152 7 -178.899 371.8 38.09 0.000
## 8 5 -183.193 376.4 42.68 0.000
## 24 6 -182.472 376.9 43.24 0.000
## 622 8 -186.379 388.8 55.05 0.000
## 4718 1.535e-05 9 -186.226 390.5 56.74 0.000
## 558 7 -188.673 391.3 57.64 0.000
## 526 6 -190.129 392.3 58.55 0.000
## 590 7 -189.871 393.7 60.03 0.000
## 14 5 -199.709 409.4 75.71 0.000
## 46 6 -199.310 410.6 76.91 0.000
## 110 7 -198.457 410.9 77.21 0.000
## 78 6 -199.554 411.1 77.40 0.000
## 38 5 -210.757 431.5 97.81 0.000
## 6 4 -213.511 435.0 101.32 0.000
## 12 5 -223.174 456.3 122.64 0.000
## 76 6 -222.249 456.5 122.79 0.000
## 268 6 -223.131 458.3 124.55 0.000
## 28 6 -223.145 458.3 124.58 0.000
## 332 7 -222.147 458.3 124.59 0.000
## 92 7 -222.230 458.5 124.75 0.000
## 284 7 -223.073 460.1 126.44 0.000
## 348 8 -222.089 460.2 126.47 0.000
## 2396 9 -221.829 461.7 127.95 0.000
## 527 6 -224.912 461.8 128.12 0.000
## 655 7 -224.848 463.7 129.99 0.000
## 783 7 -224.909 463.8 130.11 0.000
## 911 8 -224.819 465.6 131.93 0.000
## 9103 2.149e-04 9 -224.471 466.9 133.23 0.000
## 15 5 -229.371 468.7 135.03 0.000
## 143 6 -228.963 469.9 136.22 0.000
## 271 6 -229.058 470.1 136.41 0.000
## 74 5 -230.690 471.4 137.67 0.000
## 10 4 -231.707 471.4 137.71 0.000
## 399 7 -228.943 471.9 138.18 0.000
## 11 4 -233.317 474.6 140.93 0.000
## 267 5 -233.214 476.4 142.72 0.000
## 525 5 -238.877 487.8 154.05 0.000
## 9 3 -244.215 494.4 160.72 0.000
## 13 4 -243.291 494.6 160.87 0.000
## 7 4 -331.686 671.4 337.66 0.000
## 135 5 -331.231 672.5 338.75 0.000
## 5 3 -354.349 714.7 380.99 0.000
## 4 4 -397.645 803.3 469.58 0.000
## 20 5 -397.356 804.7 471.00 0.000
## 2 3 -400.251 806.5 472.79 0.000
## 3 3 -430.077 866.2 532.45 0.000
## 1 2 -435.188 874.4 540.67 0.000
## Models ranked by AIC(x)
Fixed term is "(Intercept)"
Global model call: glm(formula = Score ~ Type * NoS * ASL * AWL, family = gaussian,
data = NNS.Index.df2)
---
Model selection table
(Int) ASL AWL NoS Typ ASL:AWL ASL:NoS ASL:Typ AWL:NoS AWL:Typ NoS:Typ
624 -3.182000 0.11930 0.464700 0.0619000 0.0196000 0.004805 -5.933e-04 -3.595e-04
model.glm.2.best <- glm(Score ~ Type + NoS + ASL + AWL + ASL:NoS + ASL:Type + NoS:Type, data = NNS.Index.df2, family = gaussian)
summary(model.glm.2.best)
##
## Call:
## glm(formula = Score ~ Type + NoS + ASL + AWL + ASL:NoS + ASL:Type +
## NoS:Type, family = gaussian, data = NNS.Index.df2)
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -3.1819233 0.4886938 -6.511 2.43e-10 ***
## Type 0.0195963 0.0038642 5.071 6.26e-07 ***
## NoS 0.0619002 0.0168319 3.678 0.00027 ***
## ASL 0.1192692 0.0281447 4.238 2.85e-05 ***
## AWL 0.4647475 0.0598652 7.763 8.14e-14 ***
## NoS:ASL 0.0048050 0.0013258 3.624 0.00033 ***
## Type:ASL -0.0005933 0.0002459 -2.413 0.01633 *
## Type:NoS -0.0003594 0.0000661 -5.438 9.79e-08 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.1375812)
##
## Null deviance: 220.559 on 378 degrees of freedom
## Residual deviance: 51.043 on 371 degrees of freedom
## AIC: 333.71
##
## Number of Fisher Scoring iterations: 2
plot(allEffects(model.glm.2.best))
check_model(model.glm.2.best)
check_normality(model.glm.2.best)
## There's no formal statistical test for normality for generalized linear
## model.
## Please use `plot()` on the return value of this function:
## `plot(check_normality(model))`
plot(check_normality(model.glm.2.best))
## There's no formal statistical test for normality for generalized linear
## model.
## Please use `plot()` on the return value of this function:
## `plot(check_normality(model))`
## For confidence bands, please install `qqplotr`.
分布は、正規分布を前提 family=gaussian
ランダム効果は入れない
glm(family=gaussian) = lm()
model.lm.2.best <- lm(Score ~ Type + NoS + ASL + AWL + ASL:NoS + ASL:Type + NoS:Type, data = NNS.Index.df2)
summary(model.lm.2.best)
##
## Call:
## lm(formula = Score ~ Type + NoS + ASL + AWL + ASL:NoS + ASL:Type +
## NoS:Type, data = NNS.Index.df2)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.22198 -0.24595 -0.00821 0.27040 1.29764
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -3.1819233 0.4886938 -6.511 2.43e-10 ***
## Type 0.0195963 0.0038642 5.071 6.26e-07 ***
## NoS 0.0619002 0.0168319 3.678 0.00027 ***
## ASL 0.1192692 0.0281447 4.238 2.85e-05 ***
## AWL 0.4647475 0.0598652 7.763 8.14e-14 ***
## NoS:ASL 0.0048050 0.0013258 3.624 0.00033 ***
## Type:ASL -0.0005933 0.0002459 -2.413 0.01633 *
## Type:NoS -0.0003594 0.0000661 -5.438 9.79e-08 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.3709 on 371 degrees of freedom
## Multiple R-squared: 0.7686, Adjusted R-squared: 0.7642
## F-statistic: 176 on 7 and 371 DF, p-value: < 2.2e-16
plot(allEffects(model.lm.2.best))
check_model(model.lm.2.best)
check_normality(model.lm.2.best)
## OK: residuals appear as normally distributed (p = 0.296).
plot(check_normality(model.lm.2.best))