apply involving array and matrix

Suppose I have data, P, that is a 3x3x10 array and a matrix, q, that is a 10x3 matrix. A function, pv_fun, takes P[,,i] as an input and q[i,] as an input:

P=array(runif(90,0,1),c(3,3,10))
q=matrix(runif(30,0,1),nrow = 10)

pv.fun<-function(Prob,Prev){
  solve(diag(as.vector(Prob%*%Prev)))%*%Prob%*%diag(Prev)
}

This works if I specify "i"

pv.fun(P[,,1],q[1,])
           [,1]      [,2]       [,3]
[1,] 0.02192327 0.6626384 0.31543835
[2,] 0.52426958 0.4473938 0.02833658
[3,] 0.34857273 0.5157891 0.13563820

However, I would like to avoid a loop and use apply, but I can't figure out how.

If I try: apply(P,3,pv.fun,q)

This is close, i think: I'm telling it to use the 3rd dimension of P but how to I tell 'apply' to use the 1st dimension of q?

Thanks in advance!

Rich

Solution

How about using lapply() instead of apply(). The apply() way would work if you could use all of q for each iteration of apply, but since you need to subset P and q, that seems like a less useful strategy.

P=array(runif(90,0,1),c(3,3,10))
q=matrix(runif(30,0,1),nrow = 10)

pv.fun<-function(Prob,Prev){
  solve(diag(as.vector(Prob%*%Prev)))%*%Prob%*%diag(Prev)
} 


lapply(1:nrow(q), function(i)pv.fun(P[,,i], q[i, ]))
#> [[1]]
#>           [,1]      [,2]      [,3]
#> [1,] 0.1492837 0.2669265 0.5837899
#> [2,] 0.1502058 0.5824942 0.2673000
#> [3,] 0.1411805 0.2513680 0.6074515
#> 
#> [[2]]
#>            [,1]      [,2]       [,3]
#> [1,] 0.08494298 0.6144880 0.30056900
#> [2,] 0.03750454 0.7719693 0.19052613
#> [3,] 0.19968033 0.7864584 0.01386132
#> 
#> [[3]]
#>           [,1]      [,2]        [,3]
#> [1,] 0.4515211 0.5451417 0.003337122
#> [2,] 0.2979098 0.7018926 0.000197589
#> [3,] 0.4403383 0.5592538 0.000407891
#> 
#> [[4]]
#>           [,1]       [,2]      [,3]
#> [1,] 0.4184959 0.04292074 0.5385834
#> [2,] 0.2976478 0.05667290 0.6456793
#> [3,] 0.4141699 0.12402696 0.4618032
#> 
#> [[5]]
#>           [,1]      [,2]      [,3]
#> [1,] 0.4212735 0.1986239 0.3801026
#> [2,] 0.1233982 0.3057945 0.5708073
#> [3,] 0.5879795 0.1914091 0.2206114
#> 
#> [[6]]
#>           [,1]        [,2]      [,3]
#> [1,] 0.3336499 0.257089986 0.4092602
#> [2,] 0.6231942 0.047247936 0.3295578
#> [3,] 0.4067584 0.009297672 0.5839439
#> 
#> [[7]]
#>            [,1]      [,2]      [,3]
#> [1,] 0.05933416 0.5847652 0.3559006
#> [2,] 0.12551019 0.7238791 0.1506108
#> [3,] 0.08715145 0.6421023 0.2707463
#> 
#> [[8]]
#>           [,1]      [,2]      [,3]
#> [1,] 0.5773402 0.1198203 0.3028394
#> [2,] 0.6856290 0.1707220 0.1436490
#> [3,] 0.2166357 0.3480572 0.4353071
#> 
#> [[9]]
#>           [,1]      [,2]      [,3]
#> [1,] 0.3454159 0.5091042 0.1454799
#> [2,] 0.2020375 0.6085781 0.1893844
#> [3,] 0.3496506 0.5355888 0.1147606
#> 
#> [[10]]
#>             [,1]      [,2]       [,3]
#> [1,] 0.008277708 0.8252787 0.16644363
#> [2,] 0.017052752 0.9246549 0.05829239
#> [3,] 0.029653306 0.7013757 0.26897095

^{Created on 2023-03-02 with reprex v2.0.2}