## Files in this item

FilesDescriptionFormat

application/pdf

9624549.pdf (3MB)
(no description provided)PDF

## Description

 Title: Estimation and inference for conditionally heteroscedastic models Author(s): Zhao, Quanshui Doctoral Committee Chair(s): Portnoy, Stephen L. Department / Program: Statistics Discipline: Statistics Degree Granting Institution: University of Illinois at Urbana-Champaign Degree: Ph.D. Genre: Dissertation Subject(s): Statistics Abstract: The ordinary least squares (OLS) method is known to be efficient for linear models when the errors are homogeneous with Gaussian distributions, but troublesome with heteroscedastic or non-Gaussian errors. For the latter nonstandard case, we use the weighted quantile regression (l$\sb1$) method, gaining both robustness and efficiency, with successful applications to interval forecasting of ARCH type time series models.Dynamically changing regression parameters are another discrepancy to the ordinary linear models. By using the recursive method, the dynamically evolving parameters can be estimated. Asymptotic properties are studied for paired comparison models (a chess rating system) and dynamic ARCH models. Issue Date: 1995 Type: Text Language: English URI: http://hdl.handle.net/2142/21193 Rights Information: Copyright 1995 Zhao, Quanshui Date Available in IDEALS: 2011-05-07 Identifier in Online Catalog: AAI9624549 OCLC Identifier: (UMI)AAI9624549
﻿