This file is part of IDEAS, which uses RePEc data


[ Papers | Articles | Software | Books | Chapters | Authors | Institutions | JEL Classification | NEP reports | Search | New papers by email | Author registration | Rankings | Volunteers | FAQ | Blog | Help! ]

Stochastic Gradient versus Recursive Least Squares Learning

Author info | Abstract | Publisher info | Download info | Related research | Statistics
Author Info
Sergey Slobodyan
Anna Bogomolova,
Dmitri Kolyuzhnov
Abstract

In this paper, we perform an in—depth investigation of relative merits of two adaptive learning algorithms with constant gain, Recursive Least Squares (RLS) and Stochastic Gradient (SG), using the Phelps model of monetary policy as a testing ground. The behavior of the two learning algorithms is very different. Under the mean (averaged) RLS dynamics, the Self—Confirming Equilibrium (SCE) is stable for initial conditions in a very small region around the SCE. Large distance movements of perceived model parameters from their SCE values, or “escapes”, are observed. On the other hand, the SCE is stable under the SG mean dynamics in a large region. However, actual behavior of the SG learning algorithm is divergent for a wide range of constant gain parameters, including those that could be justified as economically meaningful. We explain the discrepancy by looking into the structure of eigenvalues and eigenvectors of the mean dynamics map under SG learning. Results of our paper hint that caution is needed when constant gain learning algorithms are used. If the mean dynamics map is stable but not contracting in every direction, and most eigenvalues of the map are close to the unit circle, the constant gain learning algorithm might diverge.

Download Info
To download:

If you experience problems downloading a file, check if you have the proper application to view it first. Information about this may be contained in the File-Format links below. In case of further problems read the IDEAS help file. Note that these files are not on the IDEAS site. Please be patient as the files may be large.

File URL: http://www.cerge-ei.cz/pdf/wp/Wp309.pdf
File Format: application/pdf
File Function:
Download Restriction: no

Publisher Info
Paper provided by The Center for Economic Research and Graduate Education - Economic Institute, Prague in its series CERGE-EI Working Papers with number wp309.

Download reference. The following formats are available: HTML, plain text, BibTeX, RIS (EndNote), ReDIF
Length:
Date of creation: Oct 2006
Date of revision:
Handle: RePEc:cer:papers:wp309

Contact details of provider:
Postal: Politickych veznu 7, Praha 1
Phone: (+420) 224 005 180
Fax: (+420) 224 005 134
Email:
Web page: http://www.cerge-ei.cz
More information through EDIRC

For technical questions regarding this item, or to correct its listing, contact: (Jana Koudelkova).

Related research
Keywords: Constant gain adaptive learning E—stability recursive least squares stochastic gradient learning.

Find related papers by JEL classification:
C62 - Mathematical and Quantitative Methods - - Mathematical Methods and Programming - - - Existence and Stability Conditions of Equilibrium
C65 - Mathematical and Quantitative Methods - - Mathematical Methods and Programming - - - Miscellaneous Mathematical Tools
D83 - Microeconomics - - Information, Knowledge, and Uncertainty - - - Search, Learning, and Information
E10 - Macroeconomics and Monetary Economics - - General Aggregative Models - - - General
E17 - Macroeconomics and Monetary Economics - - General Aggregative Models - - - Forecasting and Simulation

This paper has been announced in the following NEP Reports:

Statistics
Access and download statistics

Did you know? Over five million full texts a year are downloaded through IDEAS.

This page was last updated on 2008-10-1.


This information is provided to you by IDEAS at the Department of Economics, College of Liberal Arts and Sciences, University of Connecticut using RePEc data on a server sponsored by the Society for Economic Dynamics.