On Eliminating Inferior Regression Models.
Abstract
Consider a linear regression model with (p-1) predictor variables which is taken as the 'true' model. The goal is to select of all possible reduced models such that all inferior models (to be defined) are excluded with a guaranteed minimum probability. A procedure is proposed for which the exact evaluation of the probability of a correct decision is difficult; however, it is shown that the probability requirement can be met for sufficiently large sample size. Monte Carlo evaluation of the constant associated with the procedure and some ways to reduce the amount of computations involved in the implementation of the procedure are discussed. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 01, 1981
- Accession Number
- ADA101924
Entities
People
- Deng Yuang Huang
- S. Panchapakesan
Organizations
- Purdue University