The following is a standard, but I couldn't find a single textbook for it. Could you help me find a reference for this lemma?

Given $f,f_N \in L^2 (\Omega) \cap L^1 (\Omega)$ for some probability space $(\Omega , \mu )$, then $$\left| \mathbb{E}_{\alpha} [f] - \mathbb{E}_{\alpha} [f_N] \right| \leq \|f - f_N\|_2 \, ,$$ $$ \left| {\rm Var}(f) - {\rm Var}(f_N) \right| \leq \|f-f_N\|_2 \cdot (\|f\|_2 +\|f_N\|_2 ) \, , $$ $$ |\sigma(f)-\sigma(f_N) | \leq \frac{\|f\|_2 + \|f_N\|_2}{\sigma(f) +\sigma(f_N)} \|f-f_N\|_2 \, . $$

In other words, the $L^2$ error bounds the moment approximation error.

Note- This is somewhat overlapping with this MathOverflow post. Thanks



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.