Text Mining for User Query Analysis - A 5-Step Method for Cultural Heritage Institutions

Abstract

The recent development of Web Analytics offers new perspectives to libraries, archives and museums to improve their knowledge of user needs and behaviours. In order to dive into the mind of their end users, institutions can explore queries from a digital catalogue. However, a manual exploration demands a major time commitment and only leads to limited results. This paper explores how text mining techniques can help automate the analysis of large volumes of log files. A 5-step methodology including clustering is illustrated by a case study from the State Archives of Belgium.