Graphics Reference
In-Depth Information
Figure . . [his figure also appears in the color insert.] Spineplot of leaves brushed by response
categories with superimposed tree model. he associated tree is sketched on top foreasieridentification
of individual leaves
Due to the fixed height, it is possible to visually compare sizes of the terminal
nodes that are proportional to the width of the corresponding bar. Moreover, relative
proportions of groups are easily comparable when using highlighting or brushing.
AsamplespineplotofleavesisshowninFig. . .hedisplayeddataandmodelare
thesameasinFig. . ,aswellasthecolorbrushing.Eachbarcorrespondstoaleaf,
and the width of each bar is proportional to number of cases in that particular node.
We can clearly see relative proportions of groups within each node. In addition, it
is possible to add a simple annotation on top of the plot in the form of a dendrogram
oftherepresentedtree.Aswithtreemaps, itisadvantageous tochoosethesizeofgaps
between bars according to the depth of the split.
SPOLsaremainly usefulforcomparing groupproportionswithin terminal nodes.
hey are similar to spineplots, which allow the comparison of groups within cate-
gories of a variable. hey differ in that a “category” is in fact membership of a partic-
ular terminal node and uses different rules for gaps between bars.
Inthis section wehavediscussedseveralalternative techniques forvisualizing tree
models based on the idea of recursive partitioning. he shown methods focus on the
visualization of splits, their sequence, and the application of the model to data. One
important property of all visualization techniques presented is their applicability to
arbitrary subsets of the data. Although most illustrations used training data and the
correspondingfittedtreemodel,itisalsofeasibletovisualizetestdata instead.Where
a view of the training data highlights the adaptability of the model, the view of test
data focuses on stability and overfitting. Moreover, it is possible to compare both
views side by side.
his leads us to further important aspects of a tree model - the credibility and
quality of the splits and the entire model. In the next section we want to briefly dis-
cuss tree modelconstruction and present visualization methods that incorporate in-
formation about split quality into both existing and new plots.
Search WWH ::




Custom Search