posted by .

There is one step in a proof that I don't understand. Could someone please explain?

u = any vector in vector space S
W = finite dimensional subspace of S with orthonormal basis of vectors {v1, v2, ..., vn}

The theorem to prove is: u can be expressed exactly one way as
u = w1 + w2

w1 = orthogonal projection of u onto W
w2 = component of u orthogonal to W
w = arbitray vector in subspace W

w1 = <u,v1>*v1 + <u,v2>*v2 + ... + <u,vn>*vn
w2 = u - w1
w = k1*v1 + k2*v2 + ... + kn*vn

<w2,w> = <u - w1,w> = <u,w> - <w1,w>
<u,w> = <u,k1*v1 + k2*v2 + ... + kn*vn> = k1*<u,v1> + k2*<u,v2> + ... + kn*<u,vn>
(!!! Don't understand where this equation came from --->) <w1,w> = <u,v1>*k1 + <u,v2>*k2 + ... + <u,vn>*kn
<u,w> = <w1,w> so <w2, w> = 0

Where, did that one line come from? the textbook makes a reference to this theorem but includes no other explanation:

If S is an orthornormal basis for an n-dimensional inner product space and
us = (u1,u2,...,un)
vs = (v1,v2,...,vn)
<u,v> = u1v1 + u2v2 + ... + unvn

Can anyone follow this?

You use the fact that the inner product is bi-linear. So, you can always linearly expand. E.g.:

<c_1u_1 + c_2u_2, d_1v_1+d_2v_2> =

c_1d_1<u_1,v_1> +
c_1d_2<u_1,v_2> +
c_2d_1<u_2,v_1> +
c_2d_2<u_2,v_2> +

If you take u_i = v_i to be the i-th unit vector, then:

<u_1,v_1> = 1

<u_2,v_2> = 1

<u_1,v_2> = 0

<u_2,v_1> = 0

The vector

c_1u_1 + c_2u_2

can be denoted as



d_1v_1+d_2v_2 as


The inner product is

c_1d_1 + c_2d_2

because the "cross terms" vanish as they are proportional to inner products of orthogonal vectors.

In your problem:

w1 = <u,v1>*v1 + <u,v2>*v2 + ... + <u,vn>*vn

w = k1*v1 + k2*v2 + ... + kn*vn

and if you take the inner product between w1 and w then you get a sum of the inner products of all the combinations where you pick the i-th term from w1 and the j-th term from w, summed over i and j. But these are zero unless i = j, because <v_i,v_j> = 0 if i and j are not equal. If i and j are the same then it is 1. So, you end up with the sum of

<u,v_j>k_j from j = 1 to n.

bingo, got it! In hindsight, that's pretty simple. But, I just couldn't figure that out yesterday.

thank you so much count iblis!

Respond to this Question

First Name
School Subject
Your Answer

Similar Questions

  1. math

    If A^TA is an invertible matrix, prove that the column vectors of A are linearly independent. You know that if statement X implies statement Y then that is equivalent to Not(Y) implies Not(X). You can start by taking the column vectors …
  2. math

    Prove that if A is a symmetric n x n matrix, then A has a set of n orthonormal eigenvectors. I've read the entire page and while it's on the correct topic, …
  3. Algebra

    How would you prove this theorem: The column space of an m x n matrix A is a subspace of R^m by using this definition: A subspace of a vector space V is a subset H of V that has three properties: a) the zero vector of V is in H. b) …
  4. math

    Find an orthonormal basis for the subspace of R^3 consisting of all vectors(a, b, c) such that a+b+c = 0. The subspace is two-dimensional, so you can solve the problem by finding one vector that satisfies the equation and then by constructing …
  5. math

    A trigonmetric polynomial of order n is t(x) = c0 + c1 * cos x + c2 * cos 2x + ... + cn * cos nx + d1 * sin x + d2 * sin 2x + ... + dn * sin nx The output vector space of such a function has the vector basis: { 1, cos x, cos 2x, ..., …
  6. math

    Find the least squares approximation of x over the interval [0,1] by a polynomial of the form a + b*e^x --------------------------------------------------------- The polynomial produces an output space with two linearly independent …
  7. Vectors

    Verify using an example that Vector a + (Vector b • Vector c) is not equal to (Vector a + Vector b) • (Vector a +Vector c). (This means that addition does not distribute over the dot product.) Explain the problem that arises.
  8. calculus

    State whether or not the following statements are true. Justify your reasoning.?
  9. Vectors

    Verify using an example that Vector a + (Vector b • Vector c) is not equal to (Vector a + Vector b) • (Vector a +Vector c). Explain the problem that arises
  10. math vectors

    Use a specific example to prove that the cross product is also not associative. That is, use three specific vectors in 3-space to show that Vector a×(Vector b × Vector c) is not equal to (Vector a × Vector b) × Vector c. Can you …

More Similar Questions