Theorizing the application of Transaction-Oriented Web usage Mining on typical e-commerce Web sites

Abstract

The purpose of this research is to show the importance of theorizing the application of transaction-oriented Web usage mining methods in typical-structured e-commerce Web sites via a reference model. This paper presents findings from the analysis of a simulated case with embedded rules that are generated by three Web usage mining methods and triggered in three user-interactive scenarios. A two-phase simulation experiment was conducted to mine and collect the user data in the first phase and to test mining performance in the second phase. The design of the experiment was based on suggestions of previous critical reviews on virtual field experiment and guides for the applicability and use of Web experiments in information system research. The results showed that two out of the three Web usage mining methods deduced from the proposed reference model did increase significant transaction volumes via personalized recommendations. The paper is an original study and hopefully provides a valuable simulated case with insightful application of the reference model and corresponding data analyses.