Library: | stats |
See also: | twboxcox hinesplot boxcoxdens |
Quantlet: | boxcox | |
Description: | computes the Box-Cox transformation of x |
Usage: | y = boxcox (x, p) | |
Input: | ||
x | n x q | |
p | scalar | |
Output: | ||
y | n x q transformed data |
; loads the library stats library("stats") ; loads the library graphic library("graphic") ; reset random generator randomize(0) ; generate x(skewed) x = 0.05+0.9.*uniform(250) x = qfc(x, 2) ; create a display t = createdisplay(1,2) ; show a qq-plot of x xp = grqqn(x) show(t, 1, 1, xp) ; now transform with boxcox(p=0) y = boxcox(x, 0) ; show a qq-plot of y yp = grqqn(y) show(t, 1, 2, yp)
shows first a qq plot with the original data. The data are skewed. The second plot shows a qq-plot with the transformed data. The data are less skewed now and the distribution looks more normal.