Publikation

Case-Based User Modeling in an Experience Management System

Andreas Jedlitschka, Klaus-Dieter Althoff

In: KI - Künstliche Intelligenz, German Journal on Artificial Intelligence - Organ des Fachbereiches "Künstliche Intelligenz" der Gesellschaft für Informatik e.V. (KI) "Lernen, Modellierung und Kommunikation" 1 Seiten 46-49 BöttcherIT 2003.

Abstrakt

Experience-based continuous learning is essential for improving products, processes, and technologies in emerging as well as in established areas of business and engineering science. It can be facilitated by case-based organizational learning, meaning that relevant experience is captured in the form of cases for reuse in a corporate experience repository (case base; experience base; EB). For obvious reasons, learning from experience needs to be a permanent endeavor. Thus, an organization has to handle a “continuous stream of experience”. For this purpose, a learning organization, called “Experience Factory” (EF; Basili et al. 1994, Althoff et al. 2000), was established at Fraunhofer IESE with the Corporate Information Network (CoIN) initiative (Althoff et al. 2001). On one side, CoIN grants access to intranet-based information like organizational processes and news and, on the other side, case-based access to project management experiences. The objective of the next development step of CoIN is to provide users with valuable information/knowledge at the right time, in an adequate representation, and within the actual context (“just-intime”). In this article, we describe ongoing and planned work to enhance our existing experience management system. We introduce new strategies to capture, process, disseminate, and exchange knowledge. Case-based reasoning (CBR) is used for both knowledge modeling/retrieval/adaptation as well as for a “learning from example” based approach to user modeling. A brief outlook concludes this article.

ki03-ja.pdf (pdf, 172 KB )

Deutsches Forschungszentrum für Künstliche Intelligenz
German Research Center for Artificial Intelligence