Evaluation Graphs

The Evaluation Graph has several options to help understand the prediction error of the design. Because these plots are generated before any data is collected, a standard deviation of 1 is used to calculate relative prediction error.

The fraction of design space (FDS) graph is the default evaluation graph. It shows the standard error profile for the entire design in a single, simple curve.

The remaining graph views shown on the Graphs Tool, show how the standard error changes as a function of the factor setting’s changes. They behave the same as the model graphs views of the same type, except they show the prediction variance surface rather than the response surface.

Right-click on the graph to select “graph preferences” to adjust the labels, tick marks, axis dimensions, colors, etc.

When comparing evaluation graphs for multiple designs, make sure the graph axes have the same limits for all designs.