Generalized Stochastic Gradient Learning

Show simple item record

dc.contributor.author Evans, George W., 1949-
dc.contributor.author Honkapohja, Seppo, 1951-
dc.date.accessioned 2005-12-15T16:42:42Z
dc.date.available 2005-12-15T16:42:42Z
dc.date.issued 2005-09-19
dc.identifier.uri http://hdl.handle.net/1794/1927
dc.description 35 p. en
dc.description.abstract We study the properties of generalized stochastic gradient (GSG) learning in forwardlooking models. We examine how the conditions for stability of standard stochastic gradient (SG) learning both differ from and are related to E-stability, which governs stability under least squares learning. SG algorithms are sensitive to units of measurement and we show that there is a transformation of variables for which E-stability governs SG stability. GSG algorithms with constant gain have a deeper justification in terms of parameter drift, robustness and risk sensitivity. en
dc.format.extent 419658 bytes
dc.format.mimetype application/pdf
dc.language.iso en_US en
dc.publisher University of Oregon, Dept of Economics en
dc.relation.ispartofseries University of Oregon Economics Department Working Papers ; 2005-17 en
dc.subject Adaptive learning en
dc.subject E-stability en
dc.subject Recursive least squares en
dc.subject Robust estimation en
dc.title Generalized Stochastic Gradient Learning en
dc.type Working Paper en


Files in this item

Files Size Format View
UO-2005-17_Evans_Gradient.pdf 409.8Kb PDF View/Open

This item appears in the following Collection(s)

Show simple item record