source: orange/docs/widgets/rst/visualize/radviz.rst @ 11359:8d54e79aa135

Revision 11359:8d54e79aa135, 5.9 KB checked in by Ales Erjavec <ales.erjavec@…>, 14 months ago (diff)

Cleanup of 'Widget catalog' documentation.

Fixed rst text formating, replaced dead hardcoded reference links (now using
:ref:), etc.

1.. _Radviz:
6.. image:: ../icons/Radviz.png
8Radviz vizualization with explorative data analysis and intelligent data
9visualization enhancements.
15   - Examples (ExampleTable)
16      Input data set.
17   - Example Subset (ExampleTable)
18      A subset of data instances from Examples.
19   - Attribute Selection List
20      List of attributes to be shown in the visualization.
24   - Selected Examples (ExampleTable)
25      A subset of examples that user has manually selected from the scatterplot.
26   - Unselected Examples (ExampleTable)
27      All other examples (examples not included in the user's selection).
28   - Attribute Selection List
29      List of attributes used in the visualization.
35Radviz ([Hoffman1997]_) is a neat non-linear multi-dimensional visualization
36technique that can display data on three or more attributes in a 2-dimensional
37projection. The visualized attributes are presented as anchor points equally
38spaced around the perimeter of a unit circle. Data instances are shown as
39points inside the circle, with their positions determined by a metaphor from
40physics: each point is held in place with springs that are attached at the
41other end to the attribute anchors. The stiffness of each spring is
42proportional to the value of the corresponding attribute and the point ends up
43at the position where the spring forces are in equilibrium. Prior to
44visualization, attribute values are scaled to lie between 0 and 1. Data
45instances that are close to a set of feature anchors have higher values for
46these features than for the others.
48The snapshot shown below shows a Radviz widget with a visualization of the
49data set from functional genomics ([Brown2000]_). In this particular
50visualization the data instances are colored according to the corresponding
51class, and the visualization space is colored according to the computed class
52probability. Notice that the particular visualization very nicely separates
53the data instances of the different class, making the visualization interesting
54and potentially informative.
56.. image:: images/Radviz-Brown.png
58To gain further understanding about the placement of the data points in
59two-dimensional space, it helps to set on the :obj:`Show value lines` and
60use :obj:`Tooltips show spring values`. We also switched-off the
61:obj:`Show probabilities` to see the markings associated with data points
62better. The resulting display is shown below. From it, it should be clear that
63high values of "spo5 11" attribute (and for some data instances high values of
64"spo mid") is quite characteristic for instance of class Ribo, which at the
65same time have comparable lower value of other attributes. High values of
66heat 20 and diau f are characteristic for Resp class. See [Leban2006]_ and
67[Mramor2007]_ for further illustrations of utility of Radviz in analysis of
68this and similar data set from functional genomics. Other options in the
69:obj:`Settings` tab are quite standard. The :obj:`Point size` controls the size
70of the points that mark the data instnace. :obj:`Jittering Options` are
71especially interesting when displaying data with discrete attributes, where
72many of the data instances would overlap. Same could happen also with
73continuous attributes if many data instances use the same value of the
74attributes. :obj:`Scaling Options` can shrink or blow-up the visualization from
75its central point. From :obj:`General Graph Settings`, which mainly includes
76standard point-visualization options, let us bring to your attention
77:obj:`Show value lines` which we used in the visualization below and which
78tells the widget to annotate each data point with a set of lines, each
79corresponding with each of the attributes displayed. The length of these lines
80are proportional to the attribute values (no line if the value is minimal).
81A slider accompanying this option sets the scale in which the lines are drawn.
82:obj:`Tooltip Settings` determine which information is being displayed when the
83pointer gets over the data instance.
85.. image:: images/Radviz-Brown-Springs.png
87Just like all point-based visualizations, this widget includes tools for
88intelligent data visualization (VizRank and FreeViz, see [Leban2006]_) and
89[Demsar2007]_) and interface for explorative data analysis - selection of data
90points in visualization. Just like in :ref:`Scatter Plot` widget, intelligent
91visualization can be used to find a set of attributes that would result in an
92interesting visualization. For now, this works only with class-labeled data
93set, where interesting visualizations are those that well separate data
94instances of different class. Radviz graph above is according to this
95definition an example of a very good visualization, while the one below - where
96we show an VizRank's interface (:obj:`VizRank` button in
97:obj:`Optimization dialogs`) with a list of 5-attribute visualizations and
98their scores - is not. See documentation of :ref:`Scatter Plot` widget for
99further details on VizRank, and for description of explorative analysis
100functions (selection of data instances and zooming).
105.. [Hoffman1997] Hoffman,P.E. et al. (1997) DNA visual and analytic data mining.
106   In the Proceedings of the IEEE Visualization. Phoenix, AZ, pp. 437-441.
108.. [Brown2000] Brown, M. P., W. N. Grundy, et al. (2000).
109   "Knowledge-based analysis of microarray gene expression data by using
110   support vector machines." Proc Natl Acad Sci U S A 97(1): 262-7.
112.. [Leban2006] Leban, G., B. Zupan, et al. (2006). "VizRank: Data Visualization
113   Guided by Machine Learning." Data Mining and Knowledge Discovery 13(2):
114   119-136.
116.. [Demsar2007] Demsar J, Leban G, Zupan B. FreeViz-An intelligent multivariate
117   visualization approach to explorative analysis of biomedical data. J Biomed
118   Inform 40(6):661-71, 2007.
120.. [Mramor2007] Mramor M, Leban G, Demsar J, Zupan B. Visualization-based
121   cancer microarray data classification analysis. Bioinformatics 23(16):
122   2147-2154, 2007.
Note: See TracBrowser for help on using the repository browser.