A Framework for Web Usage Mining on Anonymous Logfile Data

In this paper we point out how important generalized views of consumer behaviour can be in real life applications. These views can be extracted by performing web usage miningon log file data from anonymous visitors. After discussing the raw data we describe how additional variables can be generated from common log files. By a combination of different data mining techniques like sequence analysis algorithms, cluster analysis, and neural networks it is possible to find out structural problems in web design, typical navigation paths of visitors who will register, or web pages visited together. We show how this information can be used to improve the important customer touch point web for anonymous visitors. We clarify this methodology by showing an example from a real life application.