Normality

#Normalization checks and tests
dataset=iris
hist(dataset$Sepal.Length,breaks=20)
boxplot(dataset$Sepal.Length)
#Min Max
tf=dataset$Sepal.Length
norm.mmtf=(tf-min(tf))/(max(tf)-min(tf))
#z-score
norm.ztf=(tf-mean(tf))/sd(tf)
norm.ztf
#decimal scaling
x=floor(log10(max(tf) )) + 1
norm.dtf=tf/10^x
# Normality tests # p <.05 NOT NORMAL 
#Shapiro is preferred and powerful. Univariate
ntest.sh=shapiro.test(dataset$Sepal.Length)
ntest.sh
#Anderson-Darling Test -  gives more weight to the tails of the distribution.
library(nortest)
ntest.ad=ad.test(dataset$Sepal.Length)
ntest.ad
#Kolmogorov-Smirnov Test. For variance 
ntest.ks=ks.test(dataset$Sepal.Length,dataset$Sepal.Length)
ntest.ks