On Modeling Learning Communities Zinayida Petrushyna, Ralf Klamma, and Milos Kravcik Advanced Community Information Systems Group RWTH Aachen University Ahornstrasse 55, 52056 Germany {petrushyna,klamma,kravcik}@dbis.rwth-aachen.de Abstract. The aim to support learning communities by Web 2.0 tech- nology leads to their investigation through the prism of learning the- ories. However, self-observation and self-modeling requires appropriate tools and skills. In this paper we focus on users of a forum platform and propose several tools that perform community detection, social net- work analysis, text mining, natural language processing, and clustering. The outcomes serve as inputs for community models that are automat- ically established using the i * information modeling approach. Stake- holders can recognize issues in their communities by visual analytics of the models. Based on this, they can refine community learning processes and community environments by retrieving new community requirements from the models. The process of model establishment was evaluated by experts in i * information modeling and the results show their acceptance of the proposed techniques. This solution enables also support of various well-known modeling approaches (like IMS Learning Design). Keywords: Learning Communities. Modeling Forum Communities. Com- munities of Practice. i * Modeling. 1 Introduction Incalculable actions of users represent a major problem of information system development process. In Web 2.0, behavior of stakeholders has a huge impact on digital media sustainability. It is a challenge to model collective influence of users and retrieve their requirements regarding social media. The problem is applicable to learning communities emerging in digital media environments. In the past, success of learning depended mostly on communities of learners [1] and their activities [2–4]. With too few learner interactions communities become static and knowledge inside the communities does not change a lot [5]. If learners find a solution to a discussed issue and no more issues exist, there are no more reasons that can encourage the learners to new actions. Learning communities have been transferred into the digital world where they have to consider social media as a new actor in their environment as well as the digital media have to react to community requirements. In our particular case and due to the abstract nature of models, we operate among others with