Knowledge Management through Ontology-Driven Integration of Disparate Knowledge Sources

Article type: Research Article

Authors: Werrmann, Jan^{a; *} | Krämer, Bernd J.^b

Affiliations: [a] Global Service & Parts, Daimler AG, Böblingen, Germany | [b] FernUniversität in Hagen, Hagen, Germany

Correspondence: [*] Corresponding author. Email: jan.werrmann@daimler.com.

Abstract: Acquiring case- or situation-dependent knowledge in real-time is one of the core challenges for modern societies. This holds especially for companies handling complex, knowledge-intensive business processes. The information owned by a company is typically scattered over disparate knowledge bases, document management systems, databases, software applications, and files maintained in the company’s intranet. Enterprise search engines promise to provide the right information to the right people in the company at the right time. An inherent problem of enterprise search systems is, however, that they ignore the complex web of application-dependent relationships spanning individual information resources. For most business cases it is not sufficient to obtain just a list of hits referring to single information pieces matching a query. The search engine should also crawl explicit and hidden relationships between information resources to collect coherent knowledge matching a query. This paper suggests a combination of enterprise search technologies and knowledge representation techniques based on ontologies and user feedback processing. This way, search processes are enabled to exploit explicit or tacit relationships among information elements stored in different resource containers to compose a meaningful answer. The prototype system we implemented provides a generic knowledge structure that models knowledge items, information sources and explicit or deduced relationships between them. This knowledge structure offers a homogeneous and coherent search space. It is deeply integrated with advanced information retrieval technologies to render search processes in large disparate knowledge bases and document management systems more effective and increase the quality of search results. The prototype also serves to capture the collective intelligence of knowledge producers and knowledge users in the form of mutable weighted relationships between knowledge items. To demonstrate the use of the approach and evaluate its innovative capabilities, the prototype system was successfully tested on a large car manufacturer’s after-sales activities including maintenance, service and repair in car workshops.

DOI: 10.3233/jid-2017-0017

Journal: Journal of Integrated Design and Process Science, vol. 21, no. 4, pp. 25-39, 2017