source: orange/orange/doc/widgets/Classify/ClassificationTreeGraph.htm @ 9399:6bbe263e8bcf

Revision 9399:6bbe263e8bcf, 7.0 KB checked in by mitar, 2 years ago (diff)

Renaming widgets catalog.

Line 
1<html>
2<head>
3<title>Classification Tree Graph</title>
4<link rel=stylesheet href="../../../style.css" type="text/css" media=screen>
5<link rel=stylesheet href="../../../style-print.css" type="text/css" media=print></link>
6</head>
7
8<body>
9
10<h1>Classification Tree Graph</h1>
11
12<img class="screenshot" src="../icons/ClassificationTreeGraph.png">
13<p>Visualization of classification tree.</p>
14
15<h2>Channels</h2>
16
17<h3>Inputs</h3>
18
19<DL class=attributes>
20<DT>Classification Tree (TreeClassifier)</DT>
21<DD>Classification tree.</DD>
22</dl>
23
24<h3>Outputs</h3>
25<DL class=attributes>
26<DT>Examples (ExampleTable)</DT>
27<DD>Attribute-valued data set associated with a classification tree
28node selected by the user.</DD>
29
30</dl>
31
32<h2>Description</h2>
33
34<p>This is a versatile widget with 2-D visualization of classification
35tree that allows user to alter the appearance of the tree and
36associate tree-based data to graphical elements of the
37visualization. In addition, user can select a node, instructing the
38widget to commit the data associated to the node to its output channel,
39thus in combination with other widget obtaining interesting schemas
40for explorative data analysis.</p>
41
42<a href="ClassificationTreeGraph.png"><img class="screenshot"
43src="ClassificationTreeGraph-S.png" alt="Classification Tree Graph
44Widget" border=0></a></center>
45
46<h3>General Tab</h3>
47
48<p>Several general parameters that affect the drawing size of the
49displayed tree may be set in the General tab. Here, user can set the
50zoom factor and define the vertical and horizontal spacing between the
51nodes. Node bubble information (like the one displayed in the snapshot
52above) may be turned off, and user may instruct the widget to truncate
53the text labels to fit provided space.</p>
54
55<h3>Tree Tab</h3>
56
57<a href="ClassificationTreeGraph-TreeTab.png"><img class="screenshot"
58src="ClassificationTreeGraph-TreeTab-S.png" alt="Classification Tree Graph
59- Tree Tab" border=0 align="left"></a>
60
61<p>In this tab the user can define the depth to which the widget will
62display the tree (may be useful when very large trees are
63induced). The edges between the nodes in the tree graph are drawn
64according to the maximum width defined in this tab. All the edges will
65be of equal width if <span class="option">No dependency</span> is chosen in the
66<span class="option">Reference for Edge Width</span> box. When <span class="option">Root node</span> is selected
67instead, the width of the edge will correspond to the proportion of
68instances in the corresponding node with respect to all the instances
69in the training data set (that is, instances in the root node). That
70is, under this selection, the edge will get thinner and thinner when
71traversing toward the bottom of the tree. Selection of <span class="option">Parent
72node</span> makes the edge width correspond to proportion of instances in
73the nodes with respect to the instances in their parent node.</p>
74
75<br clear="all">
76
77<h3>Node Tab</h3>
78
79<a href="ClassificationTreeGraph-NodeTab.png"><img class="screenshot"
80src="ClassificationTreeGraph-NodeTab-S.png" alt="Classification Tree Graph
81- Node Tab" border=0></a></center>
82
83<p>Parameters from the Node Tab govern the appearance of tree nodes
84and define what is displayed in the node box, of which the size can be
85set manually (<span class="option">Node Width</span>). Tree nodes can display two lines of
86information selected by the user (<span class="option">Show Info</span> box). In the
87snapshot the probability for the target class (republican) was
88reported in the first line in the node, and number of instances from
89the training set that reached this node in the second text
90line. Because of treatment of missing values, these numbers may not
91necessary be integers.</p>
92
93<p>The color of the node:
94<ul>
95<li>may be uniform (<span class="option">Node Color</span> set to
96<span class="option">Default</span>),</li>
97<li>correspond to number of instances in the node
98with respect to the number of instances in the root node (<span class="option">Instances
99in node</span>),</li>
100<li>may report on the probability of the majority class
101(<span class="option">Majority class probability</span>) where one would expect that the
102color intensity would be higher towards the leaves of the node,</li>
103<li>may report on probability of the target class (<span class="option">Target class
104probability</span>), with more intense colors marking the nodes where
105instances of target class are more frequent, and</li>
106<li>may report on the distribution of instances with target class,
107where the intensity of node color corresponds to proportion of the
108target class instances in the node with respect to the target class
109instances in the root node (<span class="option">Target class
110distribution</span>).</li>
111</ul>
112
113<h3>Navigation</h3>
114
115<p><span class="option">Find Root</span> aligns the position of the window such that the
116root node is displayed in the center. If user selects the node
117(left-click on the node), <span class="option">Find Selected</span> would bring the widow
118in position to display the selected node in the center. More advanced
119navigation is supported by a navigator window (click on the
120<span class="option">Navigator</span> button), which displays a complete tree with a box
121signifying the currently viewed part of the tree in the main
122window. This box can be moved around thus exposing different parts of
123the tree in the main window. The tree visualization in the navigator
124corresponds to the visualization parameters set in the main
125window. Navigator also supports bubble info box which is displayed
126when mousing over the node in the tree.</p>
127
128<a href="ClassificationTreeGraph-Navigator.gif"><img class="screenshot"
129src="ClassificationTreeGraph-Navigator-S.gif" alt="Classification Tree Graph
130- Navigator" border=0></a></center>
131
132<h2>Examples</h2>
133
134<p>Following is a simple schema where we have read the data,
135constructed the classification tree and viewed it in our tree
136viewer. Notice that if both the viewer and the widget that constructs
137the classification tree are open, any run of tree induction algorithm
138will have an immediate effect in the visualization. You can thus use
139this combination to explore how are the changes in induction algorithm
140reflected in the structure of the resulting tree.</p>
141
142<a href="ClassificationTreeGraph-SimpleSchema.gif"><img class="screenshot"
143src="ClassificationTreeGraph-SimpleSchema-S.gif" alt="Classification Tree Graph
144- Simple Schema" border=0></a></center>
145
146<p>Any click on the node of the tree would place the related data
147instances on the widget's output channel. This is explored in the
148schema below that shows this data in the data table and as a data
149subset in the scatterplot visualizer (for scatterplot to display the
150selected data properly, make sure that the tree data is passed as a
151data subset; this can simply be done by connecting scatterplot to the
152Data widget first, and connecting it to the tree visualization widget
153next).</p>
154
155<a href="ClassificationTreeGraph-Interaction.png"><img class="schema"
156src="ClassificationTreeGraph-Interaction-S.png" alt="Classification Tree Graph
157- Interaction"></a>
158
159</body>
160</html>
Note: See TracBrowser for help on using the repository browser.