Download S-PLUS 6.0 for UNIX User's Guide

Transcript
Chapter 6 Menu Graphics
Example
In this example, we use a supersmoother to graphically explore the
relationship between the fifth and sixth sensors in the sensors data
set.
1. Open the Scatter Plot dialog.
2. Type sensors in the Data Set field.
3. Select V5 as the x Axis Value and V6 as the y Axis Value.
4. Click on the Fit tab and select Supersmoother as the
Smoothing Type.
5. Click Apply to leave the dialog open.
A Graph window is created containing a plot. As in the previous
examples, you can experiment with the smoothing parameter by
varying the value in the Span field. For example, click on the Fit tab
in the open Scatter Plot dialog. By default, no span value is specified,
so it is computed internally by cross-validation. Type various values
between 0.1 and 1 in the Span field, clicking Apply each time you
choose a new value. Each time you click Apply, a new Graph
window appears that displays the updated curve. Note how the
smoothness of the fit is affected. When you are finished
experimenting, click OK to close the dialog.
Multipanel
Conditioning
In the section Grouping Variables, we plotted multiple twodimensional scatter plots on the same set of axes according to the
value of a third factor (categorical) variable. It is also possible to place
multiple scatter plots on separate axes, conditioned on the value of a
third variable. When a conditioning variable is categorical, S-PLUS
generates plots for each level. When a conditioning variable is
numeric, conditioning is automatically carried out on the sorted
unique values; each plot represents either an equal number of
observations or an equal range of values.
The Scatter Plot dialog, as well as many other dialogs in the Graph
menu, includes options for specifying conditioning variables,
arranging the plots, and labeling the panels. For additional detailed
examples on conditioning in the Graph dialogs, see the section
Visualizing Multidimensional Data.
152