Scale-up Suite Help

Fitted values and confidence intervals

Fitting > Fitting Results > Fitted values and confidence intervals |

The fitted **values** and **confidence intervals** (Conf. Int.) are reported in the table at the top of the **Fitting results** panel.

The confidence intervals are expressed as a percentage of the parameter value. In general your aim should be to get the confidence interval smaller than 10% of the parameter value. If it is greater than 100%, the fit for that parameter is unreliable (statistically insignificant). In this case it is possible that you should fix the value of the parameter (remove it from the list of parameters to fit) as your data may not contain sufficient information for this parameter to be fitted.

The confidence interval on a parameter value indicates the sensitivity of the model results to changes in the parameter. If the confidence intervals are small, this implies a large sensitivity of the model to the parameter, so there is a ‘sharp’ minimum. If the confidence intervals are large, the effect of changing the parameter on the model is small, suggesting that a range of parameter values give similar results.

Three confidence levels can be selected from the dropdown: 90%, 95% and 99%. The confidence limit is based on the assumption of a normal distribution for the errors, with 90%, 95% and 99% corresponding to 1.65, 1.96 and 2.58 times the standard error of the parameter fit respectively. The standard errors can be found in the fitting report.

A preview of the fitting report can be obtained by clicking in the Fitting toolbar.