Data Mining with Microsoft SQL Server 2008 Book Review Chapter 3

DMX stands for Data Mining Extensions, though originally was called OLE DB for Data Mining, a name from the pre-.NET days. The book recalls how people on the data mining product team used “guerilla tactics” to encourage Microsoft’s official SQL marketing department to use their preferred acronym. I guess marketing had to have something to do. I’m wondering if someone has a study proving whether DMX is a better marketing phrase than OLE DB for Data Mining. I know for sure I can track how many people will read this article based on my decision to review this book and its technology.

The chapter starts with the premise concluding that “the field is relatively immature” meaning that “there are no standard concepts of mining models, training or predictions”. I have believed that there is a difference between the research-based approach that I term machine learning, and the applied science, I call data mining (I would even use the phrase data mining engineering to be consistent with other engineering disciplines). When the authors talk about the immaturity, they refer to the potential capacity of business intelligence professionals and software architects and database developers to integrate this technology across industries and even across software vendors. By contrast, I have deep respect for the decades of proven mathematical research behind the machine learning algorithms, some of which have proven to be time-tested production-level essential elements of some of the world’s most sophisticated systems.

Continue reading “Data Mining Concepts and DMX” »