Curve Fitting

Curve fitting allows for a powerful analysis method for your imported data. To start curve fitting on a data set, press the three vertical dots on the item you want to fit to, and choose "Curve Fitting". This brings up the Curve Fitting dialog for the selected data set. An extensive list of predefined equations can be chosen from the dropdown. The corresponding mathematical formula is shown in the same card.

Each detected parameter in the formula will be shown in the sidebar. Here an initial guess can be set for the fit, as well as the minimum and maximum value for this parameter. Make sure that the initial guess is in between the bounds, and that the minimum bound is smaller than the maximum bound. If you don't want to use any bounds for the parameter, you can set the bounds to infinite using -inf and inf for the minimum and maximum bound respectively.

Below the fitting parameters, a box is shown with the resulting parameters of the fit. Apart of the final results, a margin is shown as well for each parameter as well as the sum of R² indicating the goodness of the fit. The margin can be set using the primary menu in the Curve Fitting dialog as a multiple of the standard deviation, corresponding to an uncertainty margin of 65, 95 or 99.7% for a standard deviation of 1σ, 2σ or 3σ respectively.

Please note that like all fits, that this standard deviation is mostly a measure of how well the data agrees with the corresponding model. Factors such as whether this model is actually physically correct, if there's any systemic errors in the measurements, and if the parameters are correlated to each other is not taken into account when fitting data. This is an inherent property of all curve fitting, and a common point of confusion, leading to misleading error bars on data points even in scientifically published data.

When performing a fit, there are three different fitting algorithms to choose from, these can be set from the primary menu.