| Chapter 1: | Introduction |
Data mining techniques have shown their value in analyzing specific marketing problems in recent years, especially as more and more businesses start to utilize the Internet to conduct business and as data becomes abundant. While the two research communities share the same interests, their methodologies are quite different. For example, in market basket analysis, the data mining literature often focuses on discovering frequently co-purchased items (Agrawal 1995, Brijs 2000); in contrast, marketing researchers tend to build multicategory purchase models based on consumer utility theories (Manchanda 1999), and study the relationship among items, such as their complementarity and/or substitutability. In market segmentation, distance-based clustering techniques, such as k-means (Hartigan 1975), and parametric mixture models such as Gaussian mixture models (Fraley and Raftery 1998) represent two main approaches used in marketing research; various algorithm-based clustering techniques are developed in the data mining literature.
In general, data mining research incorporates methodologies from various research communities such as statistics, machine learning, database technology, optimization and pattern recognition, and hence has a richer pool of knowledge/model representation. It focuses more on the effectiveness of problem solving, and pays greater attention to the actual performance on data. On the other hand, marketing research advocates more theory-based analysis and its theories are often built upon statistics, economics, econometrics and other social sciences.
The broader objective of the research is to study how data mining and marketing approaches can be used to study marketing problems. In particular, we want to integrate and compare methods from marketing research and data mining research. This research consists of two essays that contribute to this goal. In each of the two essays in this research, both data mining and marketing approaches are used to address selected marketing problems. Although the research stops short at integrating both types of approaches on the methodology level, it does combine both approaches in a way to effectively address the problems.