Improved Decision-making in Data Mining

This book explores model selection using decision trees based on discretized data, and the accompanying implications this has for decision-makers. Model selection is a non-trivial exercise with a large impact on decision-making. An information criterion for selecting between competing decision tree models is presented along with a method for quantifying the opportunity costs of choosing a specific decision tree. After the initial information criterion development a real world example from an American insurance company call center is presented. The example includes the development of a program to automate data discretization, decision tree creation, and decision tree selection. A decision tree is chosen and critically analyzed from a managerial decision-making point of view. Actual call center performance data is used as input data and results are identified and presented in a way that is advantageous to managers.

Eric Kyper is an assistant professor of management information systems at Lynchburg College. He graduated with a Ph.D. in MIS from the University of Rhode Island. His research focuses primarily on quantitative analysis and data mining. Dr. Lloyd received his B.S.and M.S.in Information Systems from Virginia Commonwealth University. He was awarded a Information System Ph.D. from Kent State University in 1996. Since that time he has been a professor and is currently at the University of Rhode Island.