This fact is much
more emphasized when the number of nesting levels is large. Therefore, the NLD
structure offers a richer environment than OLAP operations for the exploration
and navigation in aggregated data. To illustrate this fact, one can see that the inner
diagram in the supremum of the outer lattice (node # 1) is a projection (rollup) of
the lattice (produced from the fact table in Table 1(a)) on Govern while the immediate
predecessors of that external node (i.e., nodes #2, #3, and #4) reflect a rollup
limited to Internal. To get a rollup on the combination of Internal and Govern, we
need to look at the inner lattices within outer nodes. For example, when Internal
= 1 (see node #4), 51 enterprises have either Govern = 2 (27 cases) or Govern = 3
(24 cases), but no one of them has a bad quality of governance (i.e., Govern = 1)
like in the case when Internal takes a value equal to 2 or 3 (internal structure in
nodes #3 and #2) .
We have developed a set of tools for information and knowledge visualization: (1)
the tool called CubeViz helps highlight the salient cells in a data cube as well as
associations among cells, and (2) the module NLD allows a nested representation
of lattices. The latter is built upon our DM platform called Galicia (2004).
Conclusion
In this chapter, we have presented our solution towards a mutual collaboration between
data mining and data warehousing technologies with flexibility/interactivity
and efficiency objectives in mind.
Pages:
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506