Java Reference
In-Depth Information
process for users. This includes both
data definition language
and
data
manipulation language
for building models and scoring data.
Stepping back from specific features details or capabilities, let's
consider data mining standards in general. The JDM effort started
with several goals relative to data mining standards. The JDM expert
group sought to provide a unified concept space for data mining that
covered the core data mining objects as well as aspects of the data
mining process, which included tasks such as build, apply, test,
import, and export. This built upon many of the concepts introduced
by SQL/MM DM and CWM/DM. As data mining standards con-
tinue to evolve, each needs to look at the other data mining stan-
dards to ensure as much as possible a congruence of terminology and
capability. Moreover, each should ensure that one can fit seamlessly
with the others where features dovetail. This means that concepts
among standards should map cleanly and consistently, and where
possible, terminology should be consistent and equivalent. Features
introduced in one standard should be available in other standards as
appropriate.
An ideal situation would be for definitions of concepts and
terminology to converge to a single glossary. This is a difficult task,
if for no other reason than backward compatibility. This is also dif-
ficult due to the many different schools of thought and areas of
specialization that do not coordinate terminology. However, the
community will be well-served if across different standards within
the data mining community a unified glossary can evolve, with
mappings between similar (or exact) concepts and terms.
An even more ambitious goal is the cross-pollination of data
analytic standards in general (e.g., advanced data query, online
analytical processing [OLAP], and statistics). It is clear that data
mining is but one component of an overall analytics solution.
Business analysts need the ability to use data mining results for
designing OLAP data cubes, and use OLAP analysis to drive data
mining analysis further. Having convenient points of integration
through common objects can facilitate this for developers and
users. The standards community is still far from attaining this goal.
17.5
Summary
Data mining standards have come a long way in a relatively short
period of time. This chapter explored the various data mining
standards, their history and future, and their relationship to JDM.
Search WWH ::
Custom Search