Databases Reference
In-Depth Information
to be deployed and used. Therefore, the development and use processes are
almost disregarded in these frameworks though being embedded in an implicit
way. Consequently, it is hard to see what the most crucial success factors of
a DM project are.
2.3 Conclusions on the Considered Frameworks
With respect to foundations-oriented frameworks, some DM researchers argue
for the lack of an accepted fundamental conceptual framework or a paradigm
for DM research and consequently for the need of some consensus on the
fundamental concepts. Therefore, they try to search for some mathematical
bricks for DM. And the approaches based on granular and rough computing
present good examples of such attempts. However, others may think that
the current diversity in theoretical foundations and research methods is a
good thing and also it might be more reasonable to search for an umbrella-
framework that would cover the existing variety.
Another direction of research could lie in addressing data to be mined,
DM models, and reality views through the prism of the philosophy of science
paradigm, that includes consideration of nominalistic vs. realistic ontologi-
cal beliefs, voluntaristic vs. deterministic assumptions about the nature of
every instance constituting the observed data, subjectivist vs. objectivist ap-
proaches to model construction, ideographic vs. nomothetic view at reality;
and epistemological assumptions about how a criterion to validate knowledge
discovered can be constructed.
SPSS whitepaper [6] states that “Unless there's a method, there's mad-
ness”. It is accepted that just by pushing a button someone should not expect
useful results to appear. An industry standard to DM projects CRISP-DM
is a good initiative and a starting point directed towards the development of
DM meta-artifact (methodology to produce DM artifacts). However, in our
opinion it is just one guideline, which is in too general-level, that every DM
developer follows with or without success to some extent. Process-oriented
frameworks try to address the iterativeness and interactiveness of the DM
process. However, the development process of DM artifact and use of that
artifact are poorly emphasized.
Lin in Wu et al. [40] notices that a new successful industry (as DM) can
follow consecutive phases (1) discovering a new idea, (2) ensuring its ap-
plicability, (3) producing small-scale systems to test the market, (4) better
understanding of the new technology and (5) producing a fully scaled system.
At the present moment there are several dozens of DM systems, none of which
can be compared to the scale of a DBMS system. This fact according to Lin
indicates that we are still at the third phase with the DM area.
Further Lin in Wu et al. [40] claims that the research and development
goals of DM are quite different, since research is knowledge-oriented while
development is profit-oriented. Thus, DM research is concentrated on the de-
velopment of new algorithms or their enhancements but the DM developers in
