CODEWITHSUNDEEP
rvectorization
Xhescou
Xhescou

Reputation: 1

R: Generating random points with different mean and sd for each attribute

Is there any way in R, other than with loops, to create a matrix using different means and standard deviations for the matrix's columns?

For example, I want to create a 3x4 matrix representing 3 points with 4 attributes each, such that every column (every attribute) has its own mean and sd.

Upvotes: 0

Views: 66

Answers (2)

Hong Ooi
Hong Ooi

Reputation: 57686

Yes, you can do without loops. Take advantage of the fact that matrices in R are stored in column-major order, and replicate the mean and sd vectors to match.

means <- c(1, 10, 100, 1000)
sds <- c(0.1, 1, 10, 100)
rows <- 3
cols <- 4
m <- matrix(rnorm(rows*cols, m=rep(means, each=rows), s=rep(sds, each=rows)),
            rows, cols)
m
          [,1]      [,2]      [,3]      [,4]
[1,] 0.9993278 11.694798 105.53191  841.2182
[2,] 0.8945916  9.556729  92.90462 1212.6817
[3,] 0.9889313 10.088022 113.67009  991.2138

Upvotes: 1

Jd Baba
Jd Baba

Reputation: 6118

I am not sure what you are looking for but this might help:

matrix(c(x = rnorm(n=3,mean=0.5,1),
       y = rnorm(n=3,mean=2, 4),
       z = rnorm(n=3, mean=3.5, 10),
         a= rnorm(n=3,mean=5,12)),nrow=3,ncol=4)

           [,1]      [,2]      [,3]       [,4]
[1,]  0.7876793 2.9456827  8.082376  -3.065875
[2,] -0.9956971 4.2766553 14.178532  -5.003888
[3,]  2.1224071 0.7110594  6.744876 -11.006110

Note: you might bet different results as the data is randomly generated. Each column has its own mean and standard deviation.

Upvotes: 0

Related Questions