Abstract
Recently, clinical chemistry study suggests that Deming regression or Passing-Bablok regression should be used for seeking linear relation of two measured data instead of linear regression. The author tests these three methods at the cases of a numerical experimentation and daily rainfall estimation problem. The former shows that Deming and Passing-Bablok regressions are superior to linear regression. The latter hardly shows the superiority of any method in accuracy. Above all, considering theoretical aspects, Deming and Passing-Bablok regressions are more suitable then linear regression for analyzing two measured data.
