Figure 4. Projection on lattice vs. projection on context
2 2 Missaoui, Jatteau, Boujenoui, & Naouali
Copyright ?© 2007, Idea Group Inc. Copying or distributing in print or electronic forms without written permission of
Idea Group Inc. is prohibited.
Visualization Mechanisms
In large and complex DM output, it is crucial to provide mechanisms to (1) select
the mining space (i.e., input data), (2) limit the DM output to user??™s needs and perspectives,
and (3) browse the DM results.
Our work about data visualization concerns mainly the implementation of projection
and assembly of lattices to produce nested line diagrams and see the data mining
output (drawn either from bidimensional or multidimensional data) under different
perspectives.
Nested line diagrams (Ganter & Wille, 1999) are visualization aids which allow
the construction of a lattice L as an assembly of two or many lattices L1,...,Ln by
combining the respective Hasse diagrams of the n lattices Li into a unique complex
structure. However, neither the nodes of L nor their partial order are directly
represented. Instead, the information about them is spread over the various levels
of nesting. Figure 5 presents the nested line diagram (NLD) of the product lattice
LInternal, Govern=LInternal?—LGovern. As can be seen, the line diagram of the lattice Internal
is used as an outer frame in which the diagram of Govern is embedded.
Pages:
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504