Question

我正在进行回归，并希望突出显示具有极端残值的点。

我运行回归并将剩余列添加到我的数据框中然后我设置我要在dat $ Outlier列中显示的数据标签，并设置我希望点在dat $ Color列中的颜色

基本上，当您运行此代码时，我想将dat $ Outlier的文本显示为数据标签，并根据dat $ color

显示颜色的点

time = as.POSIXct(c("2015-05-01 10:00:00","2015-05-01 10:05:00","2015-05-01 10:10:00","2015-05-01 10:15:00","2015-05-01 10:20:00"),"GMT")
s =rnorm(5)
m= rnorm(5)
dat =data.frame(t =time, s = s, m =m)
reg = lm(s~m)
dat$resid = resid(reg)
dat$Outlier = ifelse(dat$resid > sd(dat$resid)*1, as.character(dat$t), "") 
dat$Color = ifelse(dat$resid > sd(dat$resid)*1, "red", "black") 
dat
plot(s, m)
abline(reg)

例如，如果我的dat数据框如下所示：

                    t          s           m      resid             Outlier Color
1 2015-05-01 10:00:00 -0.7141181 -0.54383561 -0.3645389                     black
2 2015-05-01 10:05:00 -1.7444731  0.09249989 -0.4226707                     black
3 2015-05-01 10:10:00 -1.1257465  0.12563139  0.2466758                     black
4 2015-05-01 10:15:00  0.6201680 -0.47515076  1.0746872 2015-05-01 10:15:00   red
5 2015-05-01 10:20:00 -0.7979108 -0.60000735 -0.5341534                     black

如何将第4行显示为2015-05-01 10:15:00作为数据标签并将该点涂成红色？

Answer 1

您可以使用points和text：

set.seed(1)

time = as.POSIXct(c("2015-05-01 10:00:00","2015-05-01 10:05:00","2015-05-01 10:10:00","2015-05-01 10:15:00","2015-05-01 10:20:00"),"GMT")
s =rnorm(5)
m= rnorm(5)
dat =data.frame(t =time, s = s, m =m)
reg = lm(s~m)
dat$resid = resid(reg)
dat$Outlier = ifelse(dat$resid > sd(dat$resid)*1, as.character(dat$t), "") 
dat$Color = ifelse(dat$resid > sd(dat$resid)*1, "red", "black") 
dat
plot(s, m, lwd=3)
abline(reg)

outlierRows <- which(dat$Outlier != "")

points(s[outlierRows],m[outlierRows],col="red",lwd=8)

text(s[outlierRows],m[outlierRows],
     labels=dat$Outlier[outlierRows],
     pos=1)

> dat
                    t          s          m       resid             Outlier
1 2015-05-01 10:00:00 -0.6264538 -0.8204684 -0.37276065                    
2 2015-05-01 10:05:00  0.1836433  0.4874291 -0.08680989                    
3 2015-05-01 10:10:00 -0.8356286  0.7383247 -1.20662950                    
4 2015-05-01 10:15:00  1.5952808  0.5757814  1.28941997 2015-05-01 10:15:00
5 2015-05-01 10:20:00  0.3295078 -0.3053884  0.37678008                    
  Color
1 black
2 black
3 black
4   red
5 black
>

Answer 2

library(ggplot2)
ggplot(data = dat) +
    geom_point(aes(x = m, y = s, colour = Color)) +
    geom_abline(slope = reg$coefficients[[2]],
                intercept = reg$coefficients[[1]]) +
    geom_text(aes(label=Outlier, x = m, y = s)) +
    scale_color_identity(guide = "none")

如何突出回归中的某些点

2 个答案: