Efficiency requires innovation

arXiv1902.06802MaRDI QIDQ6314243FDOQ6314243

Authors: Abram M. Kagan 

Publication date: 18 February 2019

Abstract: In estimation a parameter $hetainmathbbR$ from a sample $(x_1,ldots,x_n)$ from a population $P_{heta}$ a simple way of incorporating a new observation $x_{n+1}$ into an estimator $ildehet{a}_n=ildehet{a}_n(x_1,ldots,x_n)$ is transforming $ildehet{a}_n$ to what we call the {it jackknife extension} $ildehet{a}_{n+1}^{(e)}=ildehet{a}_{n+1}^{(e)}(x_1,ldots,x_n,x_{n+1})$, [ ilde heta_{n+1}^{(e)} = { ilde heta_n (x_1 ,ldots,x_n)+ ilde heta_n (x_{n+1},x_2 ,ldots,x_n) + ldots + ilde heta_n (x_1 ,ldots,x_{n-1},x_{n+1})}/(n+1).] Though $ildehet{a}_{n+1}^{(e)}$ lacks an innovation the statistician could expect from a larger data set, it is still better than $ildehet{a}_n$, [{ m var}( ilde heta_{n+1}^{(e)})leqfrac{n}{n+1} { m var}( ilde heta_n).] However, an estimator obtained by jackknife extension for all $n$ is asymptotically efficient only for samples from exponential families. For a general $P_{heta}$, asymptotically efficient estimators require innovation when a new observation is added to the data. Some examples illustrate the concept.

This page was built for publication: Efficiency requires innovation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6314243)