Quetelet’s Body Mass Index
작업환경 정리
- 현재 작업디렉토리 찾아보기
getwd()## [1] "/Users/coop2711/Dropbox/works/class/Stat_Methods/R.WD"- 검색가능한 package 와 data 열거
search()## [1] ".GlobalEnv" "package:stats" "package:graphics"
## [4] "package:grDevices" "package:utils" "package:datasets"
## [7] "KoreaEnv" "package:methods" "Autoloads"
## [10] "package:base"rn96.txt 자료 읽어들이고, 기초통계 요약하기
- rn96.txt 자료 읽어들이기, 원 자료를 보고 header 매개변수 추가
rn96<-read.table("rn96.txt",header=TRUE)- rn96 의 자료구조 살피기.
str(rn96)## 'data.frame': 41 obs. of 2 variables:
## $ height: int 161 155 158 170 160 156 162 158 158 167 ...
## $ weight: int 50 49 42 65 60 52 58 46 45 51 ...- height 와 weight 의 기초통계 살피기
summary(rn96)## height weight
## Min. :150.0 Min. :42.00
## 1st Qu.:156.0 1st Qu.:48.00
## Median :159.0 Median :52.00
## Mean :159.3 Mean :52.02
## 3rd Qu.:162.0 3rd Qu.:55.00
## Max. :170.0 Max. :65.00- 평균과 표준편차만 살펴보려면
apply(rn96,2,mean)## height weight
## 159.26829 52.02439apply(rn96,2,sd)## height weight
## 4.266289 5.667838- height 와 weight 의 산점도 그리기
plot(weight~height, data=rn96)
- 선형회귀선 추가하기
plot(weight~height, data=rn96)
abline(lm(weight~height, data=rn96)$coefficient)
- 선형모형으로 분석하기
lm(weight~height, data=rn96)##
## Call:
## lm(formula = weight ~ height, data = rn96)
##
## Coefficients:
## (Intercept) height
## 0.2369 0.3252- 회귀계수와 관련 통계량 살피기
summary(lm(weight~height, data=rn96))##
## Call:
## lm(formula = weight ~ height, data = rn96)
##
## Residuals:
## Min 1Q Median 3Q Max
## -9.6120 -3.2868 -0.5875 2.7622 13.9893
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.2369 32.8626 0.007 0.994
## height 0.3252 0.2063 1.576 0.123
##
## Residual standard error: 5.565 on 39 degrees of freedom
## Multiple R-squared: 0.0599, Adjusted R-squared: 0.0358
## F-statistic: 2.485 on 1 and 39 DF, p-value: 0.123- lowess 회귀선 추가하기
plot(weight~height, data=rn96)
abline(lm(weight~height, data=rn96)$coefficient)
lines(lowess(rn96$height,rn96$weight),col="red")
- png 포맷으로 내보내기
png("rn-plot.png",height=640,width=640)
plot(weight~height, data=rn96)
abline(lm(weight~height, data=rn96)$coefficient)
lines(lowess(rn96$height,rn96$weight),col="red")
dev.off()## quartz_off_screen
## 2BMI 계산하고 줄기-잎 그리기
- 조금 편하게 작업하기 위해서
attach(rn96)
search()## [1] ".GlobalEnv" "rn96" "package:stats"
## [4] "package:graphics" "package:grDevices" "package:utils"
## [7] "package:datasets" "KoreaEnv" "package:methods"
## [10] "Autoloads" "package:base"- BMI 공식은 \(\frac{몸무게(kg)}{키^2(m)}\)
rn96$BMI<-round(weight/(height/100)^2,digits=1)
head(rn96)## height weight BMI
## 1 161 50 19.3
## 2 155 49 20.4
## 3 158 42 16.8
## 4 170 65 22.5
## 5 160 60 23.4
## 6 156 52 21.4- 다음 작업이 왜 필요한지 생각해 볼 것.
detach(rn96)
attach(rn96)- BMI 값들의 줄기-잎 그림 그리기
stem(BMI)##
## The decimal point is at the |
##
## 16 | 855
## 18 | 00223480135555569
## 20 | 11478145667
## 22 | 1355146
## 24 | 17
## 26 |
## 28 | 0- 조금더 알아보기 쉽게 매개변수를 바꾸기
stem(BMI,scale=2)##
## The decimal point is at the |
##
## 16 | 8
## 17 | 55
## 18 | 0022348
## 19 | 0135555569
## 20 | 11478
## 21 | 145667
## 22 | 1355
## 23 | 146
## 24 | 17
## 25 |
## 26 |
## 27 |
## 28 | 0- 각 변수는정규분포에 가까울까?
qqnorm(weight)
qqnorm(height)
qqnorm(BMI)
- weight와 height의 줄기-잎 그림
stem(weight)##
## The decimal point is at the |
##
## 42 | 00
## 44 | 000
## 46 | 000
## 48 | 000000
## 50 | 00000
## 52 | 000000000
## 54 | 000
## 56 | 000
## 58 | 0
## 60 | 000
## 62 | 00
## 64 | 0stem(height)##
## The decimal point is at the |
##
## 150 | 0
## 152 | 0
## 154 | 0000000
## 156 | 00000
## 158 | 0000000
## 160 | 000000000
## 162 | 00000
## 164 | 000
## 166 | 0
## 168 | 0
## 170 | 0- Shapiro-Wilk test
shapiro.test(weight)##
## Shapiro-Wilk normality test
##
## data: weight
## W = 0.9694, p-value = 0.3304shapiro.test(height)##
## Shapiro-Wilk normality test
##
## data: height
## W = 0.9836, p-value = 0.8096shapiro.test(BMI)##
## Shapiro-Wilk normality test
##
## data: BMI
## W = 0.9466, p-value = 0.05333- Kolmogorov-Smirnov test
ks.test(weight, "pnorm", mean=mean(weight), sd=sd(weight))## Warning in ks.test(weight, "pnorm", mean = mean(weight), sd = sd(weight)):
## ties should not be present for the Kolmogorov-Smirnov test##
## One-sample Kolmogorov-Smirnov test
##
## data: weight
## D = 0.1359, p-value = 0.4355
## alternative hypothesis: two-sidedks.test(height, "pnorm", mean=mean(height), sd=sd(height))## Warning in ks.test(height, "pnorm", mean = mean(height), sd = sd(height)):
## ties should not be present for the Kolmogorov-Smirnov test##
## One-sample Kolmogorov-Smirnov test
##
## data: height
## D = 0.0904, p-value = 0.8906
## alternative hypothesis: two-sidedks.test(BMI, "pnorm", mean=mean(BMI), sd=sd(BMI))## Warning in ks.test(BMI, "pnorm", mean = mean(BMI), sd = sd(BMI)): ties
## should not be present for the Kolmogorov-Smirnov test##
## One-sample Kolmogorov-Smirnov test
##
## data: BMI
## D = 0.1195, p-value = 0.6018
## alternative hypothesis: two-sided