Keywords
[1] The projection of a vector onto the line through is the closest point
[2] The error is perpendicular to : Right triangle has
[3] The projection of onto a subspace is the cloest vector in ; is orthogonal to .
[4] is invertible (and symmetric) only if has independent columns:
[5] Then the projection of onto the column space is the vector
[6] The projection matrix onto is . It has and
Project onto a Line
|
---|
Introduction to Linear Algebra, 5th edition |
Suppose a line going through the origin. Along that line, we want to find the point closest to . The key to projection is orthogonality
The line from to is perpendicular to the vector .
Then, the projection will be some multiple of so we call .
Now we use a bit of algebra.
The projection of onto the line through is the vector
Special cases
[1] If then . Then, the projection of onto is itself:
[2] If is perpendicular to , then so the projection is
Projection Matrix
In the formula of , what matrix is multiplying ? We can decompose it into
The projection matrix
The rank of is and is in .
Projection onto a Subspace
|
---|
Introduction to Linear Algebra, 5th edition |
Suppose vectors in and these ’s are linearly independent. Now, we want to find the combination closest to a given vector .
We suppose the matrix has columns as . The combinations in are the vectors in the column space. We’re looking for the particular combination , the projection, that is closest to .
Then,
But is perpendicular to ,
Hence, we get
where is the projection of onto and is the projection matrix.
Notice that since is orthogonal to , is in the left-nullspace . Also, is invertible only if has linearly independent columns.
The projection matrix satisfies ,, and
References
[1] Introduction to Linear Algebra, 5th edition