ISSN:
1572-9974
Quelle:
Springer Online Journal Archives 1860-2000
Thema:
Informatik
,
Wirtschaftswissenschaften
Notizen:
Abstract Consistent estimation requires finding the global maximum of some specifiedobjective function. Local maxima are often easy to find, but do notnecessarily yield consistent estimates. In many nonlinear applications, theresearcher can rarely be certain that a found local maximum is global. Thispaper examines the ability of random draw tests and specification tests todetect spurious maxima. For random draw tests we analyze Veall (1990) and atest introduced here based on a generalized beta distribution. Specificationtests routinely used by researchers are also examined as methods for detectingspurious maxima. Monte Carlo results are reported on various test functions.The results suggest that specification testa are more useful than the randomdraw tests for detecting spurious maxima.
Materialart:
Digitale Medien
URL:
http://dx.doi.org/10.1023/A:1008777102696
Permalink
