Masters Oral Defense: "Nonparametric estimation of time series volatility model estimation"
Teng Tu, Washington University in Saint Louis
Abstract: In this article we consider two estimation methods of a non-parametric volatility model with autoregressive error of order two. The first estimation method based on the two-lag difference [1]. To get a better result, we consider the second approach based on the general quadratic forms [2]. For illustration, we provided several data sets from different simulation models to support the procedures of both two methods, and prove that the second approach can make a better estimation.
Keywords and phrases: autoregressive error process, two-lag difference, general quadratic forms.
Hosts: Jimin Ding & Jose Figueroa Lopez