Open chantelanuit opened 4 years ago
Maybe it would be beneficial to also add a dotted normal curve over the histogram of the residuals. In the present version, only a density curve of the data is included (that one use is useful though).
@chantelanuit the link above does not work anymore. Could you update?
@Kucharssim @fqixiang What do you think of this request? Seems valid, no?
library(MASS)
data(Animals, package="MASS") RegModel.1 <- lm(brain~body, data=Animals) summary(RegModel.1)
sresid <- studres(RegModel.1) hist(sresid, freq=FALSE, main="Distribution of Studentized Residuals") xfit<-seq(min(sresid),max(sresid),length=40) yfit<-dnorm(xfit) lines(xfit, yfit)
Fox (2002) indicates in his book:
"The standard statistics for detecting regression outliers are the studentized residuals for the model" 1
"1. Unfortunately, although the terminology that I employ is the most common, it is not universal. What I call 'studentized residuals' are sometimes termed 'externally studentized residuals' or 'deleted studentized residuals' ". (p. 192).
Thanks !