Is closeness a measure of how many entries match in matching positions? In this case:
closeness = @(x, y) nnz(x == y) / numel(x);
A = [1 2 3 4 5];
B = [1 2 3 4 5];
C = [6 7 8 9 10];
closeness(A, B)
closeness(A ,C)
If your vectors can have noninteger entries, then to mitigate against floating point errors you'd replace x == y with abs(x - y) < tol * abs(x) where tol is something small, like 1e-14.
