Personalized Structuring of Retrieved Items

People nowadays struggle with huge unstructured collections containing some important knowledge. Search engines were developed to allow an easy access to relevant information, however they themselves produce for many queries large (unstructured) result sets. We suggest to tackle this problem by structuring these result lists or even complete collections in a personalized way, i.e. as the user would have structured it. To achieve this goal, we propose to use information about structuring behavior available from, e.g, an existing bookmark or folder hierarchy of the user. Besides structuring, the final visualization should also dependent on the user’s preferences. Furthermore, we work on performance measures, which take the effect of the personalization into account to reduce the need for user studies.