Java Reference
In-Depth Information
Predictive analytics focuses more on the data mining elements of
predicting outcomes or making assignments [SearchCRM 2006], but
it has also come to mean a more automated data mining process
[Oracle-PA 2006]. Such automated data mining processes relieve the
data miner 1 of data preparation or model quality checks, allowing
the focus to be mainly on desired results (e.g., making predictions or
ranking attributes by relative importance).
Artificial Intelligence (AI) is described as “the science and
engineering of making intelligent machines, especially intelligent
computer programs. It is related to the similar task of using computers
to understand human intelligence, but AI does not have to confine
itself to methods that are biologically observable” [McCarthy 2004].
AI is very broad and dates back to at least the 1950s. As such, there
are several branches of artificial intelligence: logical AI, search, pat-
tern recognition, representation, interference, common sense knowl-
edge and reasoning, learning from experience, among others.
Aspects of data mining may find their way into each of these, but the
true emphasis is learning from experience . Machine learning [Mitchell
1997] falls into this category, where computer programs examine his-
torical data in an effort to learn or derive patterns from the data that
can be used to solve specific problems.
1.2.2
Data Mining Versus Other Forms of Advanced Analytics
When we talk to customers new to data mining and ask whether
they mine their data, we'll sometimes hear, “Yes, of course.” But
when we ask how they mine their data or which tools they use, we
hear that they have a very large database that they analyze with
complex queries. To some, this is data mining. The area of query
and reporting addresses relatively simple deductive analysis, that is,
the extraction of detail and summary data based on human-
formulated questions. For example, answering questions such as
“which stores sold portable DVD players in the past quarter” and
“how much did each sell” is common. We draw a distinction that
querying is not data mining in the machine learning sense. Answer-
ing such questions can be accomplished through a straightforward
SQL query, as shown in this code, which is based on the schema of
Figure 1-1.
1
We will sometimes refer to the user of a data mining tool or the person who
mines data as a “data miner.”
Search WWH ::




Custom Search