"Business Applications of Data Mining"
January 30, 2002
A Presentation by Mukesh Bhargava

Dr. Bhargava is a faculty member in the Dept. of Marketing and Management of the Oakland University School of Business Administration. Dr. Bhargava's presentation discussed the need for a common language between business and statistics. Only in the last 3 years the Royal Statistical Society has finally recognized the term "Data Mining". Prior to that, they wouldn't use the term. He is trying to start a center at Oakland University called CEDAR (Center for Exploratory Data Analysis & Research).

Over the last 20 years, there has been an emergence of huge data sets. The Business School is trying to learn what to do with all this data, and how to convert it into information.

A lot of money is invested in advertising. What measurements can be made of the effectiveness of the advertisement? It must be entertaining and must also sell! In Marketing they are used to working with data that has some uncertainty.

They have a lot of data of a certain kind. It is known what was bought, but they don't know why. To what extent do your past purchases predict your future purchases?

Data Mining is used for insurance premiums, targeted advertising, AT&T fraudulent call detection, Kroger, etc. The evolution of the tools has been to make inferences about people and behavior. You don't have prior hypotheses, only some common sense rules. Need decision rules that work very quickly. Strategic advantage to keeping the models secret, so not much is shared.

There is a need for multiple expertise and collaboration across Business, Statistics and Information Technology.

web sites for more info:

www.kdnuggets.com (general info)

www.acxiom.com/caseinpoint (examples)

Left: Dave Doane, Right: Mukesh Bhargava