Multicollinearity is present when the explanatory-variable columns in a multiple regression are exactly or approximately linearly dependent vectors. Exact multicollinearity makes the regression design matrix rank deficient, so the regression coefficients are not uniquely determined. Near multicollinearity can make coefficient estimates unstable and their standard errors large even when fitted values remain accurate. Common diagnostics include condition numbers, variance inflation factors, and the Farrar-Glauber test.
Multicollinearity
See also
Correlation Matrix, Farrar-Glauber Test, Linear Regression, Multiple RegressionExplore with Wolfram|Alpha
References
Belsley, D. A.; Kuh, E.; and Welsch, R. E. Regression Diagnostics: Identifying Influential Data and Sources of Collinearity. New York: Wiley, 1980.Cite this as:
Weisstein, Eric W. "Multicollinearity." From MathWorld--A Wolfram Resource. https://mathworld.wolfram.com/Multicollinearity.html