Microsoft Puts Data Mining Specification into Beta
- By Scott Bekker
Microsoft Corp. opened the beta testing phase today for its
OLE DB for Data Mining specification, a key precursor to the release of SQL
Server 2000 with its built-in data mining functionality.
OLE DB for Data Mining is an interface for integrating data
mining tools and capabilities into line-of-business and e-commerce
applications. The new protocol based on the Structured Query Language (SQL)
common to relational databases is designed to make it easier for business users
to leverage the power of data mining – previously the exclusive domain of
statistically sophisticated analysts.
The specification is similar in concept to the OLE DB for
OLAP specification released in advance of SQL Server 7.0’s release. Just as
Microsoft (www.microsoft.com) built an OLAP server into SQL Server 7.0, the company is building
data mining capabilities and interfaces code-named “Aurum,” which is Latin for
gold, into SQL Server 2000.
Industry analysts and Microsoft itself aren’t expecting
inclusion of data mining to generate as much industry buzz as OLAP did. Data
mining tools tend to be much more difficult to understand and use effectively
than OLAP tools.
Primarily the data mining functionality is expected to be
used in two ways with SQL Server 2000. Data mining tools vendors may prefer to
use a relational database to store and prepare data before running it through
data mining algorithms. Vertical ISVs may want to use the specification to
build a specific data mining algorithm into an application, such as running a
customer through an algorithm that would calculate whether the person is a good
or bad credit risk.
Vendors lining up to support the specification included
ANGOSS Software Corp., Appsource Corp., Comshare Inc., DB Miner Technology
Inc., Knosys Inc., Magnify Inc., Megaputer Intelligence Inc., Maximal
Innovative Intelligence Ltd., NCR Corp., PolyVista Inc. and SPSS Inc.
OLE DB for Data Mining was introduced in May at Tech*Ed ’99,
and has been under vendor review and modification since then under non-disclosure
The new specification includes Predictive Model Markup
Language (PMML) standards from the Data Mining Group (www.dmg.org ).
The beta specification for OLE DB for Data Mining is
available at www.microsoft.com/data/oledb.
It will be open for public review until May 15. – Scott Bekker
Scott Bekker is editor in chief of Redmond Channel Partner magazine.