I stumbled onto a linear-algebra exercise that requires me to prove that:
Y - ȳ$\vec{1}$ belongs to the orthogonal space C(1) (vector space formed by the multiples of the vector 1).
I know that I have to be able to express Y - ȳ1 as a linear combination of other vectors whose all elements are the same (e.g: α1 * [β1 β1 ... β1] + α2 * [β2 β2 ... β2] + ... + αn*[βn βn ... βn]), but I'm clueless on how to do that.
Obs.1: ȳ is the mean of the elements in vector Y, and ȳ$\vec{1}$ is the multiplication of ȳ and the all-ones vector $\vec{1}$
Obs.2: I'm not sure if the proof that I want even makes sense, so if you can please explain why if it doesn't, that'd be great.
Thanks.
The precise statement is that $y-\overline{y}(1,1,...,1):=\left(y_1-\frac{y_1+...+y_n}{n},y_2-\frac{y_1+...+y_n}{n},...,y_n-\frac{y_1+...+y_n}{n}\right)$ is orthogonal to $(1,1,...,1)$.
You can see this by multiplying $\left(y-\overline{y}(1,1,...,1)\right)\cdot (1,1,...,1)=y\cdot(1,1,...,1)-\overline{y}(1,1,...,1)=(y_1+...+y_n)-n\overline{y}=0$