Talk at WPI - Oracle In-Database Mining

My plan was to finish a full series on Oracle tools for Basic Statistics, before moving on to typical "blog-like" postings - but I realize I'll need at least another month (or two) to get to topics like ANOVA and Multiple Regression. I feel compelled to get a basic feel for a technique/concept before writing up examples of the functions here (which anybody can quickly get from the Oracle SQL Reference Manual).

So I am (reluctantly) breaking my rule to sneak in this note about a talk I gave in Nov '07 at the Worcester Polytechnic '07 Colloqium Series - before it loses its relevance. The talk provides the motivation for in-database mining, and the presentation slides offer a good intro to Oracle Data Mining in the 11gR1 Database release.

Oracle has touted the benefits of in-database mining for over 4 years now, with steady improvements to the product and an expanding customer base. The recent announcement from market leader SAS on integration with Teradata is a nice vindication of this approach. Details of the integration, whether this engineering effort pans out, and how the market receives the integrated product remains to be seen - but this is good for the industry. SAS still owns the bully pulpit in Analytics, and this move can hopefully lead to greater awareness of the benefits of this approach, and consolidation in this space. Marcos has promised to write about these developments from his perspective as a scientist/developer - so I will hold off my thoughts in anticipation of that post.

Impressions on WPI - good school with active AI and Database research groups. I had the pleasure of meeting Prof. Ruiz and Prof. Heffernan from the AI research group - they have some interesting projects there, and Prof. Rundentstein and Prof. Mani from the database research group. Check these programs out if you are an aspiring CS student.

1 comment:

ilu said...

Thanks a lot Ram for the nice info filled presentation....Keep up the good work