Mining Balanced Patterns in Web Access Data

E.H. de Graaf, J.N. Kok, and W.A. Kosters (The Netherlands)


Support Measures, Data Mining, Frequent Pattern Mining, Web Access Data


In web access analysis of a large-scale website the behaviour of visitors accessing the website is examined. An example instance of a pattern is if a visitor accesses the same parts of the website every seven days; we will call such types of patterns balanced patterns. We define balanced patterns using standard deviation and average. We propose a new approach for pruning such patterns. In comparison with related work the required algorithm and definitions will be relatively simple. Furthermore, the new pruning threshold is intuitive from an analysts perspective.

Important Links:

Go Back