A Selection Problem in Measurement Error Models
Abstract
This paper deals with a selection problem for linear measurement error models. A selection procedure is constructed, and its corresponding asymptotic optimality is also investigated. It is shown that with the assumption of the existence of the a(-th) (a>2) moment, the expected risk of the proposed selection procedure converges to 0 with the rate of order o(n(-(a/2-1))). It is further shown that when the moment generating functions of the corresponding variables exist, the expected risk of the proposed selection procedure converges with exponential order O(e(-c*n)) under mild conditions, where c* is a positive constant.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1997
- Accession Number
- ADA332193
Entities
People
- Shanti Gupta
- Xun Lin
Organizations
- Purdue University