I'm trying to take the product of the columns in a matrix using R. What am I doing wrong?-CodePudding

The problem:

The problem

What I get:

What I get

What could be the problem here? I believe I'm doing everything just like asked but still get completely different results.

CodePudding user response：

This is the result of changes (improvements) to R's default random number generator, see help("set.seed").

RNGversion('4.2.0') #current behavior
set.seed(10)
m1 <- matrix(sample(1:10, 25, replace = TRUE), nrow = 5)
apply(m1, 2, prod)
#[1] 30240 11760  8960  2520  9000

RNGversion('3.1.0') #behavior in old R versions
#Warning message:
#  In RNGkind("Mersenne-Twister", "Inversion", "Rounding") :
#  non-uniform 'Rounding' sampler used
set.seed(10)
m1 <- matrix(sample(1:10, 25, replace = TRUE), nrow = 5)
apply(m1, 2, prod)
#[1]   840   945  2016   540 10080

CodePudding user response：

The issue here is that the behaviour of set.seed was different in versions of R prior to 3.6.0. To get the expected values, you need to change the sample.kind parameter.

Old version (gives the expected values):

set.seed(10, sample.kind = "Rounding")
m1 <- matrix(sample(1:10, 25, replace = TRUE), nrow = 5, ncol = 5)
m1

     [,1] [,2] [,3] [,4] [,5]
[1,]    6    3    7    5    9
[2,]    4    3    6    1    7
[3,]    5    3    2    3    8
[4,]    7    7    6    4    4
[5,]    1    5    4    9    5

Set it back to the newer default:

set.seed(10, sample.kind = "Rejection")
m1 <- matrix(sample(1:10, 25, replace = TRUE), nrow = 5, ncol = 5)
m1
     [,1] [,2] [,3] [,4] [,5]
[1,]    9    7   10    6    9
[2,]   10    3    2    7    2
[3,]    7    8    8    6   10
[4,]    8   10    8    2    5
[5,]    6    7    7    5   10