iTelos—Case Studies in Building Domain-Specific Knowledge Graphs

The construction of a domain-specific knowledge graphs is a complex task requiring the synergistic work of domain experts, knowledge engineers, and data scientists. The goal of the iTelos methodology is to support this task when carried out by working groups with competence in computer science and in the targeted domain, but with little know-how in the software and knowledge engineering development process. The concrete usefulness of iTelos is presented by discussing six concrete case studies where it has been applied and evaluated. The main features of iTelos validated in this paper are (i) how to create an explicit and clear purpose for the construction of the domain-specific Knowledge Graph; (ii) how to enhance of the quality of the Knowledge Graph produced by defining an iterative validation of intermediate results; (iii) how to organize the project activities based on the specific roles involved; and (iv) how to make the methodology as automated and iterative as possible, resulting in a more cost-efficient and time-efficient methodology. The evaluation is both qualitative and quantitative. The results clearly show that iTelos is an important step in the right direction, still highlighting the need for refinement and improvement of the methodology.