Foundations for an expert system in domain-specific traceability

Abstract

Attempts to utilize information retrieval techniques to fully automate the creation of traceability links have been hindered by terminology mismatches between source and target artifacts. Therefore, current trace retrieval algorithms tend to produce imprecise and incomplete results. In this paper we address this mismatch by proposing an expert system which integrates a knowledge base of domain concepts and their relationships, a set of logic rules for defining relationships between artifacts based on these rules, and a process for mapping artifacts into a structure against which the rules can be applied. This paper lays down the core foundations needed to integrate an expert system into the automated tracing process. We construct a knowledge base and inference rules for part of a large industrial project in the transportation domain and empirically show that our approach significantly improves precision and recall of the generated trace links.