METHOD FOR IDENTIFYING UNCERTAINTY IN FITTING RATING CURVE WITH BAYESIAN REGRESSION

摘要

PURPOSE: An uncertainty analytical method of water level-flow rate relation curve using a Bayesian regression analysis is provided to reduce the uncertainty compared with the general regression analysis. CONSTITUTION: An uncertainty analytical method of water level-flow rate relation curve using a Bayesian regression analysis is as follows. The water level-flow rate data is actually measured or the flow rate is calculated after the water level of the constant number is randomly generated at the uniform distribution. In the water level - flow rate relation curved formula before the parameter estimation, b, that is to say the possession amount water level, is presumed by the golden ratio decomposition method. The constant number of estimation value per parameter(a,c) is gained. The parameters a and c accord to the Bayesian regression analysis of using Matlab and a statistical program based on β.