No label defined (Q3973417)

[For the entire collection see Zbl 0725.00014.] Let \(x_ 1,x_ 2,\dots,x_ n\) be i.i.d. random variables having density \(f(x,\theta_ 1,\dots,\theta_ s)\) with respect to a \(\sigma\)- finite measure \(\nu\), where \((\theta_ 1,\dots,\theta_ s)\in\Theta\subset\mathbb R^ s\). The authors consider the problem of second-order minimax risk optimal estimation of structural parameters \((\theta_ 1,\dots,\theta_ p)\) in the presence of nuisance parameters \((\theta_{p+1},\dots,\theta_ s)\), \(p<s\). Under a suitable symmetric and quasi-convex loss function, the second-order asymptotics of minimax risk are found through the solution of the first eigenvalue problem for a semi-elliptic differential operator. The differential operator becomes elliptic when the nuisance parameters are absent (i.e. \(p=s\)).