Ready-to-Use Unbiased Estimators for Multivariate Cumulants Including One That Outperforms \overline{x^3}

arXiv1904.12154MaRDI QIDQ6317852FDOQ6317852

Publication date: 27 April 2019

Abstract: We present multivariate unbiased estimators for second, third, and fourth order cumulants

C_{2} (x, y)

,

C_{3} (x, y, z)

, and

C_{4} (x, y, z, w)

. Many relevant new estimators are derived for cases where some variables are average-free or pairs of variables have a vanishing second order cumulant. The well-know Fisher k-statistics is recovered for the single variable case. The variances of several estimators are explicitly given in terms of higher order cumulants and discussed with respect to random processes that are predominately Gaussian. We surprisingly find that the frequently used third order estimator

o v e r l i n e x^{3}

for

C_{3} (x, x, x)

of a process

x

with zero average is outperformed by alternative estimators. The new (Gauss-optimal) estimator

o v e r l i n e x^{3} - 3 o v e r l i n e x^{2} o v e r l i n e x (m - 1) / (m + 1)

improves the variance by a factor of up to

5 / 2

. Similarly, the estimator

o v e r l i n e x^{2} z

for

C_{3} (x, x, z)

can be replaced by another Gauss-optimal estimator. The known estimator

o v e r l i n e x y z

for

C_{3} (x, y, z)

as well as previously known estimators for

C_{2}

and

C_{4}

of one average-free variable are shown to be Gauss-optimal. As a side result of our work we present two simple recursive formulas for finding multivariate cumulants from moments and vice versa.

This page was built for publication: Ready-to-Use Unbiased Estimators for Multivariate Cumulants Including One That Outperforms $\overline{x^3}$

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6317852)