This post continues a previous one about Microsoft SQL Server and its possibilities for mining data. DMReview, a journal dealing with data mining in practice, has an interesting article about SQL Server. In the last section of this article, Donald Farmer (the author) writes more particularly about using SQL Server for mining data. This gives a good idea of possible applications for SQL Server in the field of data mining. However, some questions are not addressed in this particular article: What about data preparation? How to choose among different algorithms for a particular data mining task? How to deal with overfitting/underfitting?