Estimating the probability density function using empirical moments is quite popular. Is there any advantage to using a different polynomial basis than the usual $1$,$x$, $x^2$ ... etc? For the moment I am interested in distributions on $\mathbb{R}+$.

If there are any advantage, are they used in practice ?

  • $\begingroup$ It may be easier to compute the coordinates in orthogonal bases than in other ones, such as $1,x,x^2,\dots$. Orthogonal bases are indeed used for density estimation; see e.g. arxiv.org/abs/1505.00275 (I don't know if that qualifies as practice, though). $\endgroup$ – Iosif Pinelis Feb 25 '18 at 22:58

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.