Contents

The explosive growth of information available online and the rapid increase of scien-tific information critically degraded the ability of scientists to keep informed about areas of interest other than their major research areas. The problem of information overload has been addressed by several approaches, such as providing personalized information by rec-ommender systems, improving search by facetted browsing environments, and facilitating the assessment of information by means of annotations. Nevertheless, none of the existing methods provides a complete and/or satisfying solution to the information overload of sci-entists, in particular mathematicians. This work therefore aims at diminishing the overload of users, which are interested in mathematical information. It is further assumed that ap-plying the economic theory of Communities of Practice (COP) by Jean Lave and Etienne Wenger [LW91] to the domain of mathematics constitutes a promising approach to address this challenge. Hence, groups of users, e.g. researchers and students, with similar tasks and practices are modeled to make predictions on the relevance as well as the user’s preferred content, structure, and appearance of a mathematical document. The user’s practices are in-ferred from his ratings, annotations, and browsing behavior. Furthermore, it is assumed that some of his practices are inscribed into the user’s collection of documents and that semantic representation formats facilitate their automated identification. Last but not least, an interac-tive, collaborative reader for mathematical documents, henceforth referred to as LECTORA, is implemented to verify this approach