Estimation of the Location of the Maximum of a Regression Function Using Extreme Order Statistics
In this paper, we consider the problem of approximating the location,x0[set membership, variant]C, of a maximum of a regresion function,[theta](x), under certain weak assumptions on[theta]. HereCis a bounded interval inR. A specific algorithm considered in this paper is as follows. Taking a random sampleX1,Â ...,Â Xnfrom a distribution overC, we have (Xi,Â Yi), whereYiis the outcome of noisy measurement of[theta](Xi). Arrange theYi's in nondecreasing order and take the average of therÂ Xi's which are associated with therlargest order statistics ofYi. This average,x0, will then be used as an estimate ofx0. The utility of such an algorithm with fixed r is evaluated in this paper. To be specific, the convergence rates ofx0tox0are derived. Those rates will depend on the right tail of the noise distribution and the shape of[theta](Â·) nearx0.
Volume (Year): 57 (1996)
Issue (Month): 2 (May)
|Contact details of provider:|| Web page: http://www.elsevier.com/wps/find/journaldescription.cws_home/622892/description#description|
|Order Information:|| Postal: http://www.elsevier.com/wps/find/supportfaq.cws_home/regional|
When requesting a correction, please mention this item's handle: RePEc:eee:jmvana:v:57:y:1996:i:2:p:191-214. See general information about how to correct material in RePEc.
If references are entirely missing, you can add them using this form.