Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have the following problem. We defined $\mathbb{H}=\{f_0,\quad f_{j,n} \quad j=1,...,2^{n-1} \quad n=1,2,...\}$ where for all $t\in[0,1]$ we put $f_0(t)=1$ and setting $K=2j-1$, $$f_{j,n}(t)=\left\{ \begin{array}{ll} 2^{\frac{n-1}{2}} & \mbox{if $x \in \big(\frac{K-1}{2^n},\frac{K}{2^n}\big)$};\\ -2^{\frac{n-1}{2}} & \mbox{if $x \in \big(\frac{K}{2^n},\frac{K+1}{2^n}\big)$};\\ 0 &\mbox{otherwise} \end{array} \right. $$

Then we are asked to show that $\mathbb{H}$ is a complete orthonormal system in $L^2[0,1]$ (Lebesgue measure). In particular $\mathbb{H}^{\perp}_{L^2[0,1]}=\{0\}$.

I have never done this before, I know what to do for orthonormal, but I can not deal with the 2 varying indices. For the completeness I have no clue what to do.

I hope anybody can help me proving this, thanks in advance!

EDIT: for orthogonal I found this: Let $\{ \psi_{j,k}(t)\}$ haar system. How to prove that it is orthogonal?

share|cite|improve this question
up vote 2 down vote accepted

There is actually a very nice way to prove completeness, which I learned in my stochastic calculus course. Just to simplify notation, let $J_{j,n}:=((j-1)2^{-n},j2^{-n}]$ for $j=1,\dots,2^n$. Then you can write $f_{n,j}(t):=2^{\frac{n}{2}}(\mathbf1_{J_{2j-1,n+1}}(t)-\mathbf1_{J_{2j,n+1}}(t))$, where $\mathbf1_{J_{2j,n+1}}(t)$ stands for the characteristic function of the set $J_{2j,n+1}$. Moreover, let


then $\mathcal{B}_n$ increase to $\mathcal{B}[0,1]$. Let $f\in L^2[0,1]$, we have by the standard example of conditional expectation:

$$E[f|\mathcal{B}_n]=\sum_{j=1}^{2^{n+1}}2^{n+1}(\int f\mathbf1_{J_{j,n+1}}d\lambda)\mathbf1_{J_{j,n+1}}\tag{1}$$

where $\lambda$ denotes the Lebesgue measure. Now suppose $(f,f_{n,j})=0$ for all $n,j$ then one obtains by induction that $\int f\mathbf1_{J_{j,n+1}}d\lambda=0$ for all $n,j$. Hence the RHS in $(1)$ vanishes. But using the martingale convergence theorem, the LHS of $(1)$ converges to $f$ $\lambda$-a.s. on $[0,1]$. So we conclude that if $f$ is orthogonal in $L^2[0,1]$ to all $f_{n,j}$ we get $f\equiv 0$.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.