You are not currently logged in.
Access JSTOR through your library or other institution:
A Comparison of Detransformed Logarithmic Regressions and Power Function Regressions
Geografiska Annaler. Series A, Physical Geography
Vol. 67, No. 1/2 (1985), pp. 61-70
Stable URL: http://www.jstor.org/stable/520466
Page Count: 10
You can always find the topics here!Topics: Power functions, Linear regression, Sediments, Regression analysis, Data lines, Mathematical independent variables, Least squares, Logarithms, Power lines, Mathematical dependent variables
Were these topics helpful?See something inaccurate? Let us know!
Select the topics that are inaccurate.
Preview not available
A comparison of detransformed logarithmic regressions and power function regressions is illustrated with regressions of sediment concentration in a Swedish river. Quite different parameters were obtained when power function equations were developed from log-transformed data compared with development from the original data. This is a consequence of different model assumptions. Power function regressions are calculated from an additive model. While detransformed logarithmic regressions represent a multiplicative model. Detransformed logarithmic regression equations are biased with respect to the original data. Some of the bias can be reduced by a correction factor.
Geografiska Annaler. Series A, Physical Geography © 1985 Swedish Society for Anthropology and Geography