Business Intelligence Network business intelligence resources

Blog: William McKnight

« McKnight Associates, Inc. makes Inc. 500 | Main | RFID Monitors Beer »

SQL Server 2005 data mining

As a couple of comments have indicated (link1, link2), the data mining capabilities of SQL Server 2005 are much improved. SQL Server 2000 had decision tree and clustering algorithms.

SQL Server 2005 adds these 5:
Time Series
Sequence Clustering
Naive Bayes
Association Rules
Neural Nets

Through the efforts of Microsoft in SQL Server 2005 as well as others before them like Polyvista, data mining is well on the way to being out of the back room with the scientists and into the hands of the analyst end-user.

Data mining is often the true analytical end result desired, yet many fear the complexities involved and do patchwork mining step-by-step, often bailing out before achieving the desired result or accepting a result with only the limited data that was able to be included without using true data mining.

Both data mining and desktop OLAP should be presented as first alternatives for end users, depending on the requirements of course. Education of end users will be key to acceptance,

  Posted by William McKnight on October 24, 2005 1:32 PM |

Comments

The white paper by PolyVista: Beyond Knowledge Discovery is an interesting one.

Thank you so much for pointing us to that site.

If you could have some comments on MS Excel 12 or Office 12, that would be great.

Posted by: Fayu Zhang | November 2, 2005 7:05 PM

Does anybody have train and test data used in SQL Server 2005 Data Mining demo by Raman Lyer?

If yes, please e-mail to dispatch@reduct.com - can be flat file.

Thanks, Adam

Posted by: Adam | July 11, 2006 4:34 PM

Post a comment