Abstracts - faqs.org

Abstracts

Mathematics

Search abstracts:
Abstracts » Mathematics

Bootstrap methods for median regression models

Article Abstract:

The problems associated with using the least-absolute deviation (LAD) estimator for a median-regression model ran through a boostrap procedure was surmounted with the development of a smoothed LAD estimator that satisfy standard conditions for obtaining asymptotic refinements through the bootstrap procedure. The smoothed LAD smooths the cusps in the solution of the standard LAD by replacing the indicator function with a smooth function, making the smoothed LAD asymptotically equivalent to the standard LAD.

Author: Horowitz, Joel L.
Publisher: Blackwell Publishers Ltd.
Publication Name: Econometrica
Subject: Mathematics
ISSN: 0012-9682
Year: 1998
Research and Development in the Physical, Engineering, and Life Sciences, Economics, Research and Development in the Social Sciences and Humanities, Statistics, Models, Economic research, Statistics (Mathematics), Statistical hypothesis testing, Approximation theory, Approximation

User Contributions:

Comment about this article or add new information about this topic:

CAPTCHA


Bootstrap critical values for tests based on generalized-method-of-moments estimators

Article Abstract:

Monte Carlo experiments are performed to evaluate the feasibility of using bootstrap critical values as generalized-method-of-moments estimators. The values' ability to handle dependent data and identify restrictions are considered in the analysis. Results show that the use of the bootstrap critical values reduces the errors in level that occur when critical values based on the first-order asymptotic theory are employed.

Author: Hall, Peter, Horowitz, Joel L.
Publisher: Blackwell Publishers Ltd.
Publication Name: Econometrica
Subject: Mathematics
ISSN: 0012-9682
Year: 1996
Methods, Monte Carlo method, Monte Carlo methods, Parameter estimation

User Contributions:

Comment about this article or add new information about this topic:

CAPTCHA


Identification and robustness with contaminated and corrupted data

Article Abstract:

Finite bounds may be placed on unidentified population parameters found in error models derived under assumptions of robust estimation. The latter, which seeks to obtain point estimators which remain unaffected by data error, can thus be replaced by bound estimation whenever the sampled data is contaminated or corrupted. This is because the bounds provide information that is otherwise not given by robust estimation.

Author: Manski, Charles F., Horowitz, Joel L.
Publisher: Blackwell Publishers Ltd.
Publication Name: Econometrica
Subject: Mathematics
ISSN: 0012-9682
Year: 1995
Statistical sampling, Estimation theory, Sampling (Statistics)

User Contributions:

Comment about this article or add new information about this topic:

CAPTCHA


Subjects list: Usage, Research, Econometrics
Similar abstracts:
  • Abstracts: On the fair division of a heterogeneous commodity. Existence of equilibrium with nonconvexities and finitely many agents
  • Abstracts: A note on Abreu-Matsushima mechanisms. Virtual implementation in iteratively undominated strategies: complete information
  • Abstracts: Strategic consensus. Competitive equilibrium with incomplete financial markets. Real indeterminacy of stationary equilibria in matching models with divisible money
  • Abstracts: Interim core concepts for a bayesian pure exchange economy. Equitable allocation of divisible goods. Market allocation of indivisible goods
  • Abstracts: Erratum to "Do pure indivisibilities prevent core equivalence? Core equivalenc theorem in an atomless economy with purely indivisible commodities only
This website is not affiliated with document authors or copyright owners. This page is provided for informational purposes only. Unintentional errors are possible.
Some parts © 2025 Advameg, Inc.