A Modified Linear, Two-Variable Regression Using Least Distance Square Method for Modelling Hydraulic/Hydrologic

摘要

In regression analysis of hydraulic and hydrological experimental data, the standard simple least square method (LSM) is generally used. The regression equation resulting from this method depends on which variable is selected as the predictand and which as the predictor. We propose a least distance square method (LDSM) whose regression equation is not variable-dependent as LSM but deriving the coefficients requires an iterative approach such as using the "Solver" program of MS-Excel. Some examples of the application of LDSM in hydraulic engineering practices are given in the paper.