Suppose $w \in L^{1}(G)$ and $w \geq 0$. The geometric mean of $w$ is defined by $$ \Delta(w) = \exp \int_{G} \log w(x) \ dx$$

where $\Delta(w) = 0$ if the integral is $-\infty$. What is the intuition behind this definition?

Source: Fourier Analysis on Groups by Walter Rudin


Isn't this an extension of the geometric mean in discrete setting?

If $a$ is the geometric mean of $a_1,a_2,a_3,\ldots,a_n$ then $$a=\sqrt[n]{a_1 \times a_2 \times \cdots \times a_n}$$ which implies $$\log(a) = \frac{\log(a_1) + \log(a_2) + \cdots + \log(a_n)}{n}$$ i.e. $\log(a)$ is the arithmetic mean of $\log(a_i)$'s.

When we talk of average in continuous sense, we replace the summation by integrals and hence

$$\displaystyle \log(\Delta(w)) = \frac{\int_G \log(w(x)) dx}{\int_G dx}$$


$$\Delta(w) = \exp \left(\frac{\int_G \log(w(x)) dx}{\int_G dx} \right)$$

and if $l \rightarrow -\infty$, then $\displaystyle \lim_{l \rightarrow -\infty} e^l = 0$ and hence $\Delta(w) = 0$ if the integral is $-\infty$


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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