Categorized Ternary Graphs - Advanced Tab

Graphical Analytic Techniques

Select the Advanced tab of the Categorized Ternary Graphs dialog to access the options described here.

Graph Type. Select the type of ternary plot to be plotted from the Graph Type list on this tab. Click the desired plot link below for a brief description of that type of graph.

Variables. Click the Variables button to display the standard variable selection dialog, in which you can select the variable(s) to be plotted and the grouping variable to be categorized for the graph. The selection that you make will then be displayed in the area of the dialog below the Variables button. Note that the selected grouping variable does not have to be a categorical variable (e.g., contain codes); you can use one of the methods of categorization options to categorize a continuous variable. (The grouping variable is not necessary if the categories are defined via the Multiple Subsets method.)

Intervals. These options allow you to choose the method of categorization for the ternary graph (i.e., for the grouping variable that will be used to categorize values of the component and response variables in each small graph; note that grouping variables do not need to be specified if Multiple Subsets categories are selected). Each of the methods is discussed in Method of Categorization.

Fit. You can choose from four regression functions to fit to the data in a ternary plot. Click a function name below for more information about that function. See also, Fitting Options for Ternary Plots. Select Off if you do not want to fit a function to the ternary plot.

Note: User-defined functions. User-defined functions can be specified in the Custom Function options pane of the Graph Options dialog. These functions are plotted over (i.e., overlaid) and not fitted to the data. In order to fit (and plot) user-defined functions to the data, use the Nonlinear Estimation module.

Data Range. Use the options in this group box to specify the data range for the ternary plots.

Ignore cases with invalid sums. Use this check box to manage cases within specified subsets (categories) in which the component proportions do not add to the constrained value. If you select this check box, STATISTICA will ignore cases in which the proportions add to a specified expected sum. If you clear this check box, then STATISTICA will rescale the proportions so that they add to 1 in each case.

Expected sum of x+y+z. Specify here the sum to which the proportions should add. Note this box is only available if the Ignore cases with invalid sums check box has been selected.