Building an Empirical Software Engineering Research Knowledge Base from Heterogeneous Data Sources

Abstract

Recently, the Systematic Knowledge Engineering (SKE) process has been introduced to help researchers build up an empirical software engineering (EMSE) Body of Knowledge (BoK) based on a systematic literature review process. However, the SKE process does not explain how to effectively capture and represent the EMSE knowledge to enable efficient data analysis. In this paper, we introduce the EMSE Research Knowledge Base Building (RKB) process, which guides knowledge engineers in developing and using a knowledge base (KB) for the SKE process based on contributions from heterogeneous data sources. We evaluate the RKB process in the context of three research topics from the EMSE domain: software inspection experiments, theory construct identification, and threats to validity. Major results are that the RKB process is effective in guiding the knowledge engineer to build a KB that allows answering the EMSE-specific queries. The RKB process shows promising results in the EMSE research context and should be investigated in other research contexts as well.