Please use this identifier to cite or link to this item: https://hdl.handle.net/10419/22612 
Year of Publication: 
2005
Series/Report no.: 
Technical Report No. 2005,21
Publisher: 
Universität Dortmund, Sonderforschungsbereich 475 - Komplexitätsreduktion in Multivariaten Datenstrukturen, Dortmund
Abstract: 
A new nonparametric estimate of a convex regression function is proposed and its stochastic properties are studied. The method starts with an unconstrained estimate of the derivative of the regression function, which is firstly isotonized and then integrated. We prove asymptotic normality of the new estimate and show that it is first order asymptotically equivalent to the initial unconstrained estimate if the regression function is in fact convex. If convexity is not present the method estimates a convex function whose derivative has the same Lp-norm as the derivative of the (non-convex) underlying regression function. The finite sample properties of the new estimate are investigated by means of a simulation study and the application of the new method is demonstrated in two data examples.
Subjects: 
nonparametric regression
order restricted inference
convexity
Nadaraya-Watson estimate
nondecreasing rearrangement
Document Type: 
Working Paper

Files in This Item:
File
Size
242.46 kB





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.