Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Interaction analysis: An algorithm for interaction prediction and activity recognition in adaptive systems

: Nazemi, Kawa; Stab, Christian; Fellner, Dieter W.


Chen, W. ; Institute of Electrical and Electronics Engineers -IEEE-:
IEEE International Conference on Intelligent Computing and Intelligent Systems, ICIS 2010. Proceedings. Vol.1 : 29-31 Oct. 2010, Xiamen, China
New York, NY: IEEE, 2010
ISBN: 978-1-4244-6583-5
ISBN: 978-1-4244-6582-8
International Conference on Intelligent Computing and Intelligent Systems (ICIS) <2, 2010, Xiamen>
Conference Paper
Fraunhofer IGD ()
user modeling; adaptive user interface; interaction analysis; user behavior; statistics; Forschungsgruppe Semantic Models, Immersive Systems (SMIS)

Predictive statistical models are used in the area of adaptive user interfaces to model user behavior and to infer user information from interaction events in an implicit and non-intrusive way. This information constitutes the basis for tailoring the user interface to the needs of the individual user. Consequently, the user analysis process should model the user with information, which can be used in various systems to recognize user activities, intentions and roles to accomplish an adequate adaptation to the given user and his current task.
In this paper we present the improved prediction algorithm KO*/19, which is able to recognize, beside interaction predictions, behavioral patterns for recognizing user activities. By means of this extension, the evaluation shows that the KO*/19-Algorithm improves the Mean Prediction Rank more than 19% compared to other well-established prediction algorithms.