CODEWITHSUNDEEP
rmatrixvarnegative-number
Runner Bean
Runner Bean

Reputation: 5155

R language, var matrix gives negative values

So when using the R programming language, (which I'm very new to) I have a 4x2 matrix "gg" and when I do var(gg) it returns a 2x2 matrix as follows. 

> gg <- matrix(c(4,2,4,5, 9, 14, 2, 32), nrow = 4, ncol = 2, byrow = TRUE)
> gg
     [,1] [,2]
[1,]    4    2
[2,]    4    5
[3,]    9   14
[4,]    2   32

> var(gg)
           [,1]   [,2]
[1,]   8.916667 -11.25
[2,] -11.250000 182.25

now I understand that the 8.916 and the 182.25 are the variances of the first and second columns.

But what is the -11.5 ?

I have read that it could be a computed covariance or correlation? But a covariance cant be negative and a correlation must be between -1 and 1 so -11.5 satisfies neither of those two conditions. So I really don't understand what this -11.5 is representing?

Upvotes: 0

Views: 1012

Answers (1)

Andrew Garcia
Andrew Garcia

Reputation: 9

var(gg) where gg is a matrix of columns x and y returns:

cov(xx), cov(xy)
cov(yx), cov(yy)

cov(xy) = cov(yx) because cov is defined as
E((x - xbar)(y - ybar)) = E((y - ybar)(x - xbar))

Upvotes: 1

Related Questions