apply different probabilities to different matrix columns

[katbak]

Hi!
I'm very new to modelling in R. I need to insert 0 and 1 into an empty matrix based on a probability. The probabilities differ for every 2 columns.
for example:

''''''r
#my empty matrix
PopulationA = matrix(0,35,8)
probability of infection
Infprob <- c(0.01, 0.50, 0.09, 0.4)
'''''''

I know how to apply a probability to the whole matrix using sapply (shown below). But I'm not able to apply different values to different columns for a binomial result.

'''''r

applying probability to whole matrix

prevalence <- 0.01
Pregnant <- sapply(15:50 ,function(p) rbinom(35000,1,prevalence))
'''''''''''

Thank you!

[technocrat]

It can help to apply a little abstraction, borrowing from school algebra.

f(x) = y

where

x is some object or object at hand
y is some object that is to be achieved
f is some function or function to transform x into y

Here

a <- matrix(0,35,8) # a matrix of zeros of dim 35,8
v <- c(0.01, 0.50, 0.09,0.4) # a vector of typeof numeric of length 4

I've abstracted away the name of the matrix from PopulationA to just a because what the matrix represents in terms of its relationship to the real world just doesn't matter. Likewise, for v.

y will also be a matrix of equal dim with a. In it will be replaced all 0 values with values of v "for every 2 columns"

a <- matrix(0,35,8)
v <- c(0.01, 0.50, 0.09, 0.4)
v <- rep(v, each = 2)

numbers <- 1:dim(a)[2]

vectors <- split(numbers, ceiling(seq_along(numbers) / 2))

runner <- vectors[1:4]

for(i in seq_along(runner)) a[,runner[i][[1]]] = v[runner[i][[1]]]

a |> head()
#>      [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8]
#> [1,] 0.01 0.01  0.5  0.5 0.09 0.09  0.4  0.4
#> [2,] 0.01 0.01  0.5  0.5 0.09 0.09  0.4  0.4
#> [3,] 0.01 0.01  0.5  0.5 0.09 0.09  0.4  0.4
#> [4,] 0.01 0.01  0.5  0.5 0.09 0.09  0.4  0.4
#> [5,] 0.01 0.01  0.5  0.5 0.09 0.09  0.4  0.4
#> [6,] 0.01 0.01  0.5  0.5 0.09 0.09  0.4  0.4

^{Created on 2023-06-27 with reprex v2.0.2}

Because the columns/variables are unnamed, the order in which the elements of v are inserted into m doesn't matter—they can be rearranged later. That is, instead of this column layout

#> [1] 0.01 0.01  0.5  0.5 0.09 0.09  0.4  0.4

we wanted

#> [1] 0.01 0.50 0.09 0.40 0.01 0.50 0.09 0.40

we could just

m <- m[,c(1,3,5,7,2,4,6,8]

or we could redefine how we construct v

v <- c(0.01, 0.50, 0.09, 0.4)
v <- c(v,v)
v
#> [1] 0.01 0.50 0.09 0.40 0.01 0.50 0.09 0.40

Also, if this is a one off, we can do numbers more directly as

numbers = 1:8

By working both ends against the middle, we come up with the successive transforms that make up the abstract f. The key insight is that the original Infprob vector needed to be expanded to match the number of columns of the original PopulationA.