Simple Regression Procedure for Power Function

Abstract

A direct identification technique based on linear weighted regression analysis is presented to estimate the power function parameters. The computations required are simple, and a hand-held calculator might be sufficient. Parameters of the power function are obtained when the function is fit to published infiltration and soil moisture–pressure head data by: (1) The standard log-linear technique; (2) the proposed linear weighted regression; and (3) a nonlinear least-squares procedure. The described method was found to be superior to the log-linear technique. Athough the linear weighted least-squares method does not fit the experimental data as well as the nonlinear regression, the simplicity and rapidity of the described method compared to a nonlinear procedure makes it an attractice alternative for estimating the power function parameters.