|On this page…|
The following types of plots are available for viewing and evaluating the estimation:
Cost function — Plot the cost function value as it changes during estimation.
Measured and simulated — Plot empirical data against simulated data.
Parameter trajectory — Plot the parameter values as they change.
Residuals — Plot the error between the experimental data and the simulated output.
When you run estimation, Parameter Estimation tool automatically displays a parameter trajectory plot that shows the change in the parameter values by iteration. However, you can create other plots for viewing the progress of the estimation before you begin estimating the parameters.
You can plot the measured data by clicking Add Plot on the Parameter Estimation tab and selecting the experiment to use for estimation under Experiment Plots of the drop down list. You can then add the simulated response on top of the measured data plot by clicking Plot Model Response on the Parameter Estimation tab. Another way of plotting the measured and simulated data is to right-click the experiment name, for this example, NewData, and select Plot measured & simulated data from the list. The plot looks like this:
You can see that the measured and simulated data do not match. You can edit the labels, adjust the limits, change the units, and the font style of the plot in the Property Editor. To launch the editor, right-click the experiment plot and select Properties from the list.
You can also add a plot for the expected estimation cost by clicking Add Plot on the Parameter Estimation tab and selecting Estimation Cost from the list. You can close the message "There is no data for ExpCost, run the optimization to update the plot", by clicking the left arrow toggle to the left of the message. The plot looks like this:
When you right-click the estimated parameters or expected cost plots, you can choose to add the scaled values or save iteration data from the list.
Add the progress plots by clicking the Add Plot button on the Parameter Estimation tab and selecting the plot in the list. Selecting Parameter Trajectory adds the following plot to the tool.
Use the View tab of the Parameter Estimation tool to arrange the layout of the plots, so that Experiment plot: NewData, Iteration plot 1, and Iteration plot 2 are all visible.
When you perform the estimation, all plots update automatically.