Publication: Automatic Differentiation and Interval Arithmetic for Estimation of Econometric Functions
Introduction
Applications
Tools
Research Groups
Workshops
Publications
   List Publications
   Advanced Search
   Info
   Add Publications
My Account
About
Impress

Automatic Differentiation and Interval Arithmetic for Estimation of Econometric Functions

- incollection -
 

Author(s)
Max E. Jerrell

Published in
Computational Differentiation: Techniques, Applications, and Tools

Editor(s)
Martin Berz, Christian Bischof, George Corliss, Andreas Griewank

Year
1996

Publisher
SIAM

Abstract
Relatively simple economic models often lead to complex estimation problems. Even when the model itself is linear, the estimation problem can be nonlinear, with an unknown number of stationary points. Disequilibrium models have proved to be particularly difficult to estimate. Newer global optimization techniques, such as simulated annealing and genetic algorithms, have not been particularly successful in solving such models. More traditional methods can fail completely unless good starting points are used. The gradient vectors and Hessian matrices of nonlinear econometric functions tend to be complex and lead to well known user and computational approximation errors. Automatic differentiation can eliminate user error. Combined with interval arithmetic, it offers a particularly useful method of global optimization.

Cross-References
Berz1996CDT

BibTeX
@INCOLLECTION{
         Jerrell1996ADa,
       author = "Max E. Jerrell",
       editor = "Martin Berz and Christian Bischof and George Corliss and Andreas Griewank",
       title = "Automatic Differentiation and Interval Arithmetic for Estimation of Econometric
         Functions",
       booktitle = "Computational Differentiation: Techniques, Applications, and Tools",
       pages = "265--272",
       publisher = "SIAM",
       address = "Philadelphia, PA",
       key = "Jerrell1996ADa",
       crossref = "Berz1996CDT",
       abstract = "Relatively simple economic models often lead to complex estimation problems. Even
         when the model itself is linear, the estimation problem can be nonlinear, with an unknown number of
         stationary points. Disequilibrium models have proved to be particularly difficult to estimate. Newer
         global optimization techniques, such as simulated annealing and genetic algorithms, have not been
         particularly successful in solving such models. More traditional methods can fail completely unless
         good starting points are used. The gradient vectors and Hessian matrices of nonlinear econometric
         functions tend to be complex and lead to well known user and computational approximation errors.
         Automatic differentiation can eliminate user error. Combined with interval arithmetic, it offers a
         particularly useful method of global optimization.",
       keywords = "interval computations, global optimization, econometrics, disequilibrium models",
       year = "1996"
}


back
  

Contact:
autodiff.org
Username:
Password:
(lost password)