Improved Decision-making in Data Mining

This book explores model selection using decisiontrees based on discretized data, and the accompanyingimplications this has for decision-makers. Modelselection is a non-trivial exercise with a largeimpact on decision-making. An information criterionfor selecting between competing decision tree modelsis presented along with a method for quantifying theopportunity costs of choosing a specific decisiontree. After the initial information criteriondevelopment a real world example from an Americaninsurance company call center is presented. Theexample includes the development of a program toautomate data discretization, decision tree creation,and decision tree selection. A decision tree ischosen and critically analyzed from a managerialdecision-making point of view. Actual call centerperformance data is used as input data and resultsare identified and presented in a way that isadvantageous to managers.