On maintaining semantic networks: challenges, algorithms, use cases

The purpose of this paper is to provide methods and algorithms to maintain a semantic network (SN). In previous work, the authors introduced the SN approach for bridging the gap of aligning enterprise information with business processes, i.e. for discovering explicit relations between them. What has been neglected so far, however, is SN maintenance, which is required to keep an SN consistent, complete and up-to-date.

The paper illustrates an approach for SN maintenance. Specifically, the authors show how an SN evolves over time, classify properties of objects and relations captured in an SN and show how these properties can be maintained. An empirical evaluation, which is based on synthetic and real-world data, investigates the performance, scalability and practicability of the proposed algorithms.

The authors prove the feasibility of the introduced algorithms in terms of runtime performance with a proof-of-concept implementation. Further, a real-world case from the automotive domain confirms the applicability of the SN maintenance approach.

As opposed to existing work, the presented approach allows for the automated and consistent maintenance of SNs. Furthermore, the applicability of the presented SN maintenance approach is validated in the context of a real-world scenario as well as two business cases.