1
$\begingroup$

Consider the unit interval $I$ with a continuous probability measure $\mu$, and consider a smooth random variable $f:I\to \mathbb{R}$. We can define its cumulative distribution function and probability density function in the regular way. There is a number of known way to obtain the latter two functions using histograms, kernel smoothing and so on. Both methods rely on the statistically motivated assumption that we randomly sample $f$.

Assume now that we have sampled $f$ on the Gauss-Legendre quadrature point $\lbrace \alpha_{i}^N \rbrace_{i=1}^N$, which are the roots of orthogonal polynomials w.r.t. $\mu$, $p_n (\alpha)$. We can now obtain an $N$-th order approximation of $f$ using $p_n$, $$ f \approx \sum\limits_{n=0}^{N-1} p_n (\alpha) \hat{f}(n) \, .$$

My Question: Is there an elegant way of approximating the CDF and/or PDF of $f$ only with the spectral coefficients $\hat{f} (n)$ and the value at tha quadrature point $f(\alpha _j ^N )$?

Note that generally speaking, $f$ is not the same as its PDF.

$\endgroup$

1 Answer 1

1
$\begingroup$

Gauss quadrature approximations of probability distributions undergird some quite new numerical methods for approximating the solution of SDEs of McKean type; see Section 2 of https://arxiv.org/abs/1608.06741 and references therein. In such SDEs, the coefficients depend on both the current state of the process and the distribution of the solution.

$\endgroup$
3
  • 1
    $\begingroup$ Nawaf, In section 2 they approximate $\phi$ with GL quadrature, but they don't construct its pdf or cdf, do they? $\endgroup$
    – Amir Sagiv
    Aug 29, 2016 at 16:23
  • $\begingroup$ I would arg that what they do is precisely what you want, since they show how to approximate $\mu(\phi)$ with GL quadrature, not $\phi$ itself. Also their case is a bit more involved since they have to propagate the weights and quadrature points in their GL approximation under the dynamics of the SDE of McKean type. They propose a clever procedure to do this based on the Euler-Maruyama method. $\endgroup$ Aug 29, 2016 at 17:52
  • $\begingroup$ PS: that is why I recommended this paper to you in the first place, since it seemed to be an "elegant" approach by necessity. $\endgroup$ Aug 29, 2016 at 18:00

Your Answer

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

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