Duke University: MGRECON 491 - Data Mining

Course Description

Data mining has been a core competency at a large number of firms for some time. It is beginning to be used as a strategic weapon that has allowed some firms to achieve industry dominance. Examples are Amazon and Harrahs. See Thomas H. Davenport, "Competing on Analytics," Harvard Business Review, January 2006.

This course teaches the concepts and skills that allow one to advise clients as a consultant, manage a data mining team, or be an effective member of a data mining team.

The prerequisite is a familiarity with regression as taught in the Fuqua core or in a typical undergraduate statistics course. Students who have exempted out of the core statistics course meet this requirement. There is, however, the issue of aptitude. This is a course in statistical methods. Students for whom statistics has been a struggle are discouraged from enrolling.

The course is organized as follows: The first three topics are (1) an overview of the subject, (2) a quick introduction to the main tools, and (3) an introduction to and a graphical illustration of the key concept of model selection via validation samples. The remainder of the course is comprised of cases from marketing, finance, and information technology with tools described and demonstrated within cases.