We used two indices: the standard deviation of relative difference (SDRD) and the mean absolute bias error (MABE).
Moreover, with a view to show performance analysis of models, the statistical testing methods such as mean absolute percentage errors (MAPE), mean absolute bias error (MABE), root mean square error (RMSE) were used.
Among the methods for estimating the field mean moisture content using the representative locations, i.e. minimal SDRD (standard deviation of relative differences), ITS (index of temporal stability), MABE (mean absolute bias error), and RMSE (root mean square error), the minimal SDRD with a constant offset provided the best results.
To indicate the performance of the models, the following statistical test methods are used: the coefficient of determination (R2), mean bias error (MBE), mean absolute bias error (MABE), mean percent error (MPE), mean absolute percent error (MAPE), root mean square error (RMSE) and the t-statistic method (tsta).
In order to evaluate the day by day performance of these models, a statistical analysis was performed by using several statistical indicators of mean absolute bias error, root mean square error, normal root mean square error, test statistic, standard deviation and coefficient of determination.
For Bandar Abbass station, the obtained mean absolute bias error (MABE) and correlation coefficient (R) for the ELM model at different depths are in the range of 0.9116 1.5988 °C and 0.9023 0.9840, respectively while for the SaE-ELM model they are in the range of 0.8660 1.5338 °C and 0.9084 0.9893, respectively.
Model evaluation for diameter fit and prediction performance was based on the global statistics mean bias (MB), mean absolute bias (MAB) and root mean square error (RMSE).
Mean bias as well as mean absolute bias was calculated.
Stepwise regression-feedforward neural network hybrid model outperformed the other models with root mean square error, mean absolute error, mean bias error and correlation coefficient values of 2.74, 2.09, 0.01 and 0.932 respectively.
The accuracy of the each model was evaluated based on the correlation coefficient, mean absolute error, mean bias error and root mean square error values.
As described previously (Marshall et al. 2008), model measurement comparisons for NO at monitoring station locations indicate reasonable to good agreement (mean bias, absolute bias, and error, 29%, 42%, and 1.6 μg/m, respectively; model measurement correlation, 0.7).
