Information Retrieval from Knowledge Management Systems: Using Knowledge Hierarchies to Overcome Keyword Limitations

Information retrieval (IR) is a key component of knowledge management systems (KMSs). KMSs frequently rely on keyword searches as a primary mechanism for retrieval. While keyword searches are very helpful to knowledge workers, they have their limitations. To illustrate some of the limitations of keyword searches, a database containing all of the articles that have appeared in MIS Quarterly was constructed. Prior research shows that ambiguous, poorly constructed keyword phrases lead to poor information retrieval results. Knowledge hierarchies have been used to overcome some of the keyword searches limitations. This research proposes the use of dimensional modeling and multidimensional database technologies to implement knowledge hierarchies. It suggests that many of the limitations inherent in keyword searches can be eliminated from knowledge management systems by exploiting the benefits of the hierarchical structure that underlies multidimensional databases.