Computes the variance of an array along a given dimension. The general syntax for its use is
y = var(x,d)
where x is an n-dimensions array of numerical type.
The output is of the same numerical type as the input. The argument
d is optional, and denotes the dimension along which to take
the variance. The output y is the same size as x, except
that it is singular along the mean direction. So, for example,
if x is a 3 x 3 x 4 array, and we compute the mean along
dimension d=2, then the output is of size 3 x 1 x 4.
The output is computed via
where
If
d is omitted, then the mean is taken along the
first non-singleton dimension of x.
The following piece of code demonstrates various uses of the var function
--> A = [5,1,3;3,2,1;0,3,1]
A =
<int32> - size: [3 3]
Columns 1 to 3
5 1 3
3 2 1
0 3 1
We start by calling var without a dimension argument, in which
case it defaults to the first nonsingular dimension (in this case,
along the columns or d = 1).
--> var(A)
ans =
<double> - size: [1 3]
Columns 1 to 2
6.333333333333333 1.000000000000000
Columns 3 to 3
1.333333333333333
Next, we take the variance along the rows.
--> var(A,2)
ans =
<double> - size: [3 1]
Columns 1 to 1
4.000000000000000
1.000000000000000
2.333333333333333