Full Title: Explaining the GetClusterCharacteristics Stored Procedure for Sequence Clustering Models

Introduction: Beginning with SQL Server 2005 Analysis Services, Microsoft introduced a new range of modeling algorithms to be used for data mining, including Sequence Clustering. Along with these algorithms, Microsoft also provides viewers that allow you to see exactly what is going on inside the models. If you run SQL Server Profiler while using the viewers, you will discover that a number of stored procedures are being called behind the scenes. In this article, I take a closer look at one of those procedures, GetClusterCharacteristics, as used against a Sequence Clustering model. This procedure provides a lot of useful information about a clustering model that I can leverage to better understand the data and my model.

NOTE: This is a whitepaper. We're including it here in addition to the Web and community-editable Wiki page for four reasons: (1) So you can print it. (2) To see the original version. (3) Read it offline. (4) To customize it for personal use. (5) To share it offline.

The community-editable, Web version of this document is available on TechNet Wiki here: http://social.technet.microsoft.com/wiki/contents/articles/970.aspx. Please visit it there if you have any information to include. You can also leave comments on the Wiki article or ask a question on this page (see "Q and A" tab).

Article by Allan Mitchell, SQL Server MVP and principal consultant at Konesans Limited in the UK. You can find Allan’s website at www.SQLIS.com.