# Matrix representation

• May 17th 2008, 07:16 AM
matty888
Matrix representation
Compute the standard matrix representation of the orthogonal projection of R^4 onto the subspace spanned by {(1,1,1,1),(1,2,3,4)}.
• May 17th 2008, 08:41 AM
Opalg
Quote:

Originally Posted by matty888
Compute the standard matrix representation of the orthogonal projection of R^4 onto the subspace spanned by {(1,1,1,1),(1,2,3,4)}.

Step 1. Find an orthonormal basis for the subspace, e.g. $\textstyle\{\frac12(1,1,1,1),\,\frac1{\sqrt{20}}(-3,-1,1,3)\}.$

Step 2. In R^n, the projection onto the one-dimensional subspace spanned by a single unit vector $(a_1,a_2,\ldots,a_n)$ has $a_ia_j$ as its (i,j)-entry. So for example the projection onto the subspace spanned by $\textstyle\frac1{\sqrt{20}}(-3,-1,1,3)$ is $\frac1{20}\begin{bmatrix}9&3&-3&-9\\ 3&1&-1&-3\\ -3&-1&1&3\\ -9&-3&3&9\end{bmatrix}$. Now do the same thing for the projection onto the subspace spanned by ½(1,1,1,1).

Step 3. If you have an orthonormal set of vectors then the projection onto the subspace spanned by them is the sum of the projections onto the one-dimensional subspaces spanned by the basis vectors. So the answer to the question is the sum of the two matrices from the previous paragraph.