왕초보 R : kertosis, skewness

[sl100104]

R은 보고입니다. Google에 가셔서 "cran R skewness"를 입력하시면 되구요. 

아래 code를 그냥 복사해서 돌려보세요.R의 무한함에 감사할 것입니다.

Part 1은 수동으로, Part2는 자동으로 계산된 결과이구요, 선물로 멋진 graph도 드립니다.

열공하세요.

###Part1

y<-rnorm(100)

hist(y, prob=T)

mykurtosis <- function(x) {  

m4 <- mean((x-mean(x))^4) 

kurt <- m4/(sd(x)^4)-3  

kurt

}

mykurtosis(y)

myskewness <-  function(x) {

m3 <- mean((x-mean(x))^3)

skew <- m3/(sd(x)^3)

skew

}

myskewness(y)

install.packages("fBasics")

library(fBasics)

kurtosis(y)

skewness(y)

hist(y) 

#source http://cran.r-project.org/doc/contrib/Ricci-distributions-en.pdf

#####Part2

x <- rnorm( 1000 , 

mean = sample( c(0, 3), size = 1000, prob = c(.4, .6),  replace = TRUE ) ) 

par( las = 1, mar = c(3,3,4,1)+.1 )  

h <- hist( x, plot = FALSE, breaks = 15 )

d <- density( x ) 

plot( h , border = NA, freq = FALSE, xlab = "", ylab = "", 

main= "Use of clipping and translucency" ) 

usr <- par( "usr" )

ncolors <- 100

dy <- ( usr[4] - usr[3] ) / ncolors

colors <- colorRampPalette( c("yellow","orange","red") )(ncolors)

abline( h = axTicks(2) , col = "gray", lwd = .5 )

for( i in 1:ncolors){

  clip( usr[1], usr[2], usr[3] + (i-1) * dy, usr[3] + i*dy )

  plot( h, add = TRUE, axes = FALSE, ylab = "", xlab = "", 

   col = colors[i], border = NA, freq = FALSE)

}

do.call( clip, as.list( usr) )

plot( h, add = TRUE, lwd = .5 , freq = FALSE, xlab = "", 

ylab = "", axes = FALSE )

lines( d, lwd = 4, col = "#22222288" )

rug( x, col = "#00000088" )

box()

kurtosis(x)

skewness(x)

#source http://addictedtor.free.fr/graphiques/RGraphGallery.php?graph=151