Measures of Lack of Fit, and Transformation of Predictors in Composite Designs.

Abstract

The central composite designs are investigated in terms of their ability to detect lack of fit and it is shown that measures of non-quadraticity can be obtained in all the k major directions. The geometrical interpretation of this is explained. The relationship of these lack of fit measures to the estimation of power transformations in the predictor variables is also explored. (Author)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
May 01, 1980
Accession Number
ADA086562

Entities

People

  • George E. P. Box
  • N. R. Draper

Organizations

  • University of Wisconsin–Madison

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Coefficients
  • Composite Materials
  • Contracts
  • Contrast
  • Data Science
  • Equations
  • Experimental Design
  • Factorial Design
  • Information Science
  • Mathematical Analysis
  • Mathematics
  • North Carolina
  • Statistical Analysis
  • Statistics
  • United States
  • Universities
  • Wisconsin

Fields of Study

  • Mathematics

Readers

  • Regression Analysis.