Putting context into knowledge management (information retrieval)
Kristian J. Hammond, Kurt D. Fenstermacher · 1999
Many information systems treat documents as word and phrases, while ignoring the context in which documents are created and accessed. However, this contextual information is critical to building more sophisticated information systems that are needed to manage knowledge within organizations. Previous systems have gathered contextual information only haphazardly, and have required manual annotation by users. People are often loath to annotate their own work, which in turn leads to either poor annotation, or none at all. By capturing the contextual information automatically, however, information systems can both reduce the burden of annotation, while also improving the quality of recommendations to users. Automatic annotation can draw on many factors, but the processes that people create to govern work are a powerful lever for understanding documents produced and accessed. These ideas evolved through the development of the SINGLESOURCE System, which was built to improve retrieval for professionals engaged in information-finding tasks. The work describes the development of SINGLESOURCE, and the larger question of applying these same techniques to other knowledge management tasks.