Architecture-based integration of CBR-components into KM-systems

Abstract

Knowledge-Management (KM)-systems integrate multiple technologies with the purpose of supporting users who have "knowledge-related" problems. Case-Based Reasoning (CBR)-technology is well known for improving problem solutions. Combining both concepts the project "TO-KNOW" has the goal to design a CBR-component that is adaptable to typical KM-Systems. The basic idea is to take into account user-queries as (intermediate) problem-solutions for getting the needed information from the KM-System's discovery services. In this paper we describe an architecture-based approach to design the CBR-component yielding a light-weighted integration into a KM-system. This is accomplished by looking at both parts designated for integration in terms of their architecture, i.e. the modularity of data and functionality. Since different KM-systems will allow various degrees of synergy, the process should result in a highly modularized CBR-component being ready for different models of integration. An overview about our design-method is given as well as examples covering first results on a more detailed level.