You are not currently logged in.
Access JSTOR through your library or other institution:
Information-Theoretic Measures of Fit for Univariate and Multivariate Linear Regressions
Henri Theil and Ching-Fan Chung
The American Statistician
Vol. 42, No. 4 (Nov., 1988), pp. 249-252
Stable URL: http://www.jstor.org/stable/2685131
Page Count: 4
Preview not available
For the purpose of measuring the relative importance of independent variables in a multiple regression, Kruskal (1987) proposed an averaging procedure over all possible orderings of these variables. The present article uses this suggestion, but it is based on a different measure, from statistical information theory, and it extends the result to systems of equations.
The American Statistician © 1988 American Statistical Association