I am trying to find the dot product of two matrices in R. In the q matrix, which must be transposed, I have three different q values that I randomly generated earlier, and in the z matrix three randomly generated z values that serve as coordinates of a random point i. I have:
z0= NULL for (i in 1:100){ z0[i]= 1 } z1= runif(100, min=0, max= 20) z2= runif(100, min=0, max=20) q0= runif(1, 0, 1) q1= runif(1, 0, 1) q2= runif(1, 0, 1) i= runif(1, 1, 101) i= ceiling(i-1) q= matrix(c(q0,q1,q2), ncol=3) z= matrix(c(z0[i],z1[i],z2[i]), ncol=3) s[i]= t(q)*z However, when I try to calculate s[i], I get Error in t(q) * z : non-conformable arrays. I am not sure why this would be as I they seem to both have the same length.
This is my first time using R so I am not really sure what is going on.
Thanks!
174 Answers
Without using matrices or any special libraries:
The dot product of two vectors can be calulated by multiplying them element-wise with * then summing the result.
a <- c(1,2,3) b <- c(4,5,6) sum(a*b) As Pascal says, dot product in R is %*%. I am able to use this successfully on your sample data:
> z0= NULL > for (i in 1:100){ + z0[i]= 1 + } > z1= runif(100, min=0, max= 20) > z2= runif(100, min=0, max=20) > q0= runif(1, 0, 1) > q1= runif(1, 0, 1) > q2= runif(1, 0, 1) > i= runif(1, 1, 101) > i= ceiling(i-1) > q= matrix(c(q0,q1,q2), ncol=3) > z= matrix(c(z0[i],z1[i],z2[i]), ncol=3) > t(q)%*%z [,1] [,2] [,3] [1,] 0.3597998 3.227388 2.960053 [2,] 0.3544622 3.179510 2.916141 [3,] 0.3550781 3.185035 2.921208 > z%*%t(q) [,1] [1,] 4.340265 Sample Answer:
library(geometry) dot(A,B) Since it seems like others have tackled your issue, I'll just add on to say that if you want a special dot product function, you can write one yourself:
dot <- function(x, y){ # x and y can be vectors or matrices result <- t(x)%*%y # %*% is the matrix multiplication operator print(result) # t(x) denotes the transpose of x } Or, as @user3503711 says in his answer, you can just use the dot() function from the geometry library.
