Evaluation of the predictive performance of biased regression estimators

David J. Friedman, Douglas Montgomery

Research output: Contribution to journalArticle

12 Citations (Scopus)

Abstract

Regression models are widely used in forecasting, either directly as prediction equations, or indirectly as the basis of other procedures. The predictive performance of a regression model can be adversely affected by both multicollinearity and high‐leverage data points. Although biased estimation procedures have been proposed as an alternative to least squares, there has been little analysis of the predictive performance of the resulting equations. This paper discusses the predictive performance of various biased estimators, emphasizing the concept that the predictive region, as well as the strength of the multicollinearity, dictates the choice of appropriate coefficient estimators.

Original languageEnglish (US)
Pages (from-to)153-163
Number of pages11
JournalJournal of Forecasting
Volume4
Issue number2
DOIs
StatePublished - 1985
Externally publishedYes

Fingerprint

Regression Estimator
Biased
Multicollinearity
Regression Model
Evaluation
Biased Estimation
Estimator
Least Squares
Forecasting
Prediction
Alternatives
Coefficient
Regression model

Keywords

  • Biased estimation
  • Multicollinearity
  • Prediction Regression analysis

ASJC Scopus subject areas

  • Modeling and Simulation
  • Computer Science Applications
  • Strategy and Management
  • Statistics, Probability and Uncertainty
  • Management Science and Operations Research

Cite this

Evaluation of the predictive performance of biased regression estimators. / Friedman, David J.; Montgomery, Douglas.

In: Journal of Forecasting, Vol. 4, No. 2, 1985, p. 153-163.

Research output: Contribution to journalArticle

@article{bc8b43266a3945d8bb430ccae2af8f72,
title = "Evaluation of the predictive performance of biased regression estimators",
abstract = "Regression models are widely used in forecasting, either directly as prediction equations, or indirectly as the basis of other procedures. The predictive performance of a regression model can be adversely affected by both multicollinearity and high‐leverage data points. Although biased estimation procedures have been proposed as an alternative to least squares, there has been little analysis of the predictive performance of the resulting equations. This paper discusses the predictive performance of various biased estimators, emphasizing the concept that the predictive region, as well as the strength of the multicollinearity, dictates the choice of appropriate coefficient estimators.",
keywords = "Biased estimation, Multicollinearity, Prediction Regression analysis",
author = "Friedman, {David J.} and Douglas Montgomery",
year = "1985",
doi = "10.1002/for.3980040205",
language = "English (US)",
volume = "4",
pages = "153--163",
journal = "Journal of Forecasting",
issn = "0277-6693",
publisher = "John Wiley and Sons Ltd",
number = "2",

}

TY - JOUR

T1 - Evaluation of the predictive performance of biased regression estimators

AU - Friedman, David J.

AU - Montgomery, Douglas

PY - 1985

Y1 - 1985

N2 - Regression models are widely used in forecasting, either directly as prediction equations, or indirectly as the basis of other procedures. The predictive performance of a regression model can be adversely affected by both multicollinearity and high‐leverage data points. Although biased estimation procedures have been proposed as an alternative to least squares, there has been little analysis of the predictive performance of the resulting equations. This paper discusses the predictive performance of various biased estimators, emphasizing the concept that the predictive region, as well as the strength of the multicollinearity, dictates the choice of appropriate coefficient estimators.

AB - Regression models are widely used in forecasting, either directly as prediction equations, or indirectly as the basis of other procedures. The predictive performance of a regression model can be adversely affected by both multicollinearity and high‐leverage data points. Although biased estimation procedures have been proposed as an alternative to least squares, there has been little analysis of the predictive performance of the resulting equations. This paper discusses the predictive performance of various biased estimators, emphasizing the concept that the predictive region, as well as the strength of the multicollinearity, dictates the choice of appropriate coefficient estimators.

KW - Biased estimation

KW - Multicollinearity

KW - Prediction Regression analysis

UR - http://www.scopus.com/inward/record.url?scp=84984503173&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84984503173&partnerID=8YFLogxK

U2 - 10.1002/for.3980040205

DO - 10.1002/for.3980040205

M3 - Article

VL - 4

SP - 153

EP - 163

JO - Journal of Forecasting

JF - Journal of Forecasting

SN - 0277-6693

IS - 2

ER -