Expert Finding for Requirements Engineering
Matthieu Vergne · Unitn-eprints PhD (University of Trento) · 2016
Requirements Engineering (RE) revolves around requirements, from their discovery to their satisfaction, passing through their formalisation, modification, and traceability with other project artefacts, like preliminary interviews or resulting source codes. Although it is clear for many that involving knowledgeable people is an important aspect of many RE tasks, no proper focus has been given to Expert Finding (EF) systems, leading to have only few related works in the field. Our work attempts to fill this gap by investigating several dimensions of EF: conceptual by analysing the literature about expertise and its evaluation, formal by revising the usual representation of expert rankings, and practical by designing an EF system. As a result, we provide (i) a metamodel grounded in literature from Psychology to identify requirements for EF systems, (ii) a novel formalisation of experts rankings which solves limitations observed in usual EF measures, (iii) two variants of an EF system which builds on usual RE indicators (accessible knowledge and social recognition), and (iv) an enriched evaluation process which investigates deeper the consistency and correctness of an EF system.