On the measures of multicollinearity in least squares regression


Journal Article

For a general regression model y = Xβ + e, E(e) = 0, Cov(e) = σ2V-1, some results on the relationship between two measures of multicollinearity, the eigenvalues and the condition numbers of X′X and X′VX, are obtained. These results are useful in examining the effects of augmentation of data on multicollinearity and the influence of an observation on the condition number of X′X in regression diagnostics. © 1990.

Full Text

Duke Authors

Cited Authors

  • Wang, SG; Tse, SK; Chow, SC

Published Date

  • January 1, 1990

Published In

Volume / Issue

  • 9 / 4

Start / End Page

  • 347 - 355

International Standard Serial Number (ISSN)

  • 0167-7152

Digital Object Identifier (DOI)

  • 10.1016/0167-7152(90)90145-W

Citation Source

  • Scopus