Modal volatility function (Q6962064)

The authors propose a non-parametric estimator for the volatility function within a broad context that encompasses nonlinear time series models as a special case. Also, they present a new and interesting result that the asymptotic properties of the proposed modal volatility estimator for stationary \(\alpha\)-mixing dependent samples are equivalent to those for independent samples under certain mild conditions. This equivalence is inherent in non-parametric estimation for dependent samples, as discussed in [\textit{Z. Cai} and \textit{E. Ould-Saïd}, Stat. Probab. Lett. 65, No. 4, 433--449 (2003; Zbl 1116.62390)]. A similar conclusion was reached by \textit{P. Hall} et al. [Ann. Stat. 23, No. 6, 2241--2263 (1995; Zbl 0854.62039)] in the context of density estimation. Additionally, they propose a data-based bandwidth selection method to assist in the practical determination of bandwidths.