Outliers are points that are distant from remaining observations.

  They can potentially skew or bias any analysis performed on the dataset.

   It is therefore very important to detect and adequately deal with outliers.

  R Package : outliers

   R Function : outlier(x,opposite =) --

   Extreme most observation from the mean.

X -- Data Frame

Opposite -- logical value, fetches outlier from the other side.

R Function : scores(x,type= chi-sq,z,t )

Type -- Normalized Scores based on type.

#Resolving Outlier

#Detecting Outlier

#install.packages(“outliers”)

library(“outliers”)

#Input

vect<-mtcars$mpg

x<-data.frame(vect)

print(x)

#Add some dummy outlier to input data

vect<-c(47,1,56,0.0272)

y<-data.frame(vect)

my_data<-rbind(x,y)

print(my_data)

#Outlier

outlier(my_data)

#Outlier from other side

outlier(my_data,opposite = TRUE)

#Visualizing Outlier using Boxplot

boxplot(my_data,notch=TRUE,col=”green”,main=”BoxPlot with Outlier”)

#Chi-Squared Scores

scores(my_data,type = “chisq”)

#Beyond 95th Percentile based on Chi-Square

scores(my_data,type = “chisq”,prob = 0.95)

#Replacing Outlier with NA values

remove_outliers<-function(x,na.rm=TRUE) {

qnt<-quantile(x, probs=c(.25, .75))

H<-1.5 * IQR(x, na.rm = na.rm)

y<-x

y[x < (qnt[1] – H)]<-NA

y[x > (qnt[2] + H)]<-NA

y

}

r<-remove_outliers(my_data$vect)

print(as.data.frame(r))

print(my_data)

#Box Plot witout Outlier

boxplot(r,notch=TRUE,col=”green”,main=”Box Plot without Outliers”)