Local \(M\)-estimation for conditional variance function with dependent data (Q289728)

The authors are interested in the nonparametric robust estimation of the conditional variance function in heteroscedastic regression models under stationary \(\alpha\)-mixing dependent samples. Actually \(M\)-type estimators of the conditional variance function and its first-order derivative are developed. 11 assumptions are given in order to describe the set of necessary conditions for the asymptotic properties of the local \(M\)-estimators. The weak consistency and the asymptotic normality of the estimators are proven. It is to be noticed that a big work of bibliography is developed in the introduction and in the conclusion of this paper.