Data Smoothing: Thin-Plate Spline

Plots Tab

You can use the Plots tab to create plots that graphically display results of the Thin-Plate Spline analysis. (See Figure 19.3.) The raw residuals are computed as $Y - \hat{Y}$, where $\hat{Y}$ indicates the variable that contains the predicted values of the response.

Creating a plot often adds one or more variables to the data table. The following plots are available:

Observed vs. Explanatory with smoother


creates a scatter plot of the X and Y variables, overlaid with a smoother.

Confidence limits for means


specifies whether to add 95% upper and lower confidence curves to the Observed vs. Explanatory plot. The meaning of the curves is described in the section "Computational Formulas" in the TPSPLINE documentation.

Observed vs. Predicted


creates a scatter plot of the Y variable versus the predicted values.

Raw residuals vs. Predicted


creates a scatter plot of the residuals versus the predicted values.

Raw residuals vs. Explanatory


creates a scatter plot of the residuals versus the X variable.

Residual normal QQ


creates a normal Q-Q plot of the residuals.

GCV vs. log(n*lambda)


creates a scatter plot of the GCV criterion versus the smoothing parameter value for a range of smoothing parameter values.

Minimum log(n*lambda)


specifies the minimum value of the smoothing parameter to consider.

Maximum log(n*lambda)


specifies the maximum value of the smoothing parameter to consider.

Number of subdivisions


specifies the number of smoothing parameters to consider. The value in this field is combined with the values in the previous two fields to form a list of values for the LOGNLAMBDA= option.

Note: SAS/IML Studio adds a smoother to an existing scatter plot when both of the following conditions are satisfied:

  • The scatter plot is the active window when you select the analysis.

  • The scatter plot variables match the analysis variables.

Chapter 18: Data Smoothing: Loess, discusses how to display multiple smoothers in a single scatter plot and how to remove smoothers from a scatter plot.