Semantic mining of competences based on contributions in an online-community
By werthner - Posted on May 28th, 2010
In many organizations people need to identify experts that are competent in certain problem domains. Competence management systems are useful tools to keep track of who has competences in specific areas. The key challenge of such systems is to determine validated competences. The main goal of SeCoMine is the development of a novel semantic approach to measure competences. In the last years social software systems became very popular. These systems support individuals to easily share their knowledge and experience. Within this project, we will investigate howcompetences of individuals can be measured based on their contributions in such social software systems. Currently, there exist no comparable approaches to quantify competences. Therefore, SeCoMine represents a clear innovation in the field of research as well as commercial solutions.SeCoMine addresses the following issues:- Adapt an existing ontology to the needs of SeCoMine. This new ontology represents the relationsbbetween competences in a specific application context. The types of concepts and relations base on a competence concept declared at the beginning of this project.- Extract relevant information from a user¿s contribution by means of semantic text mining technologies. In addition, we incorporate annotated information given by the user, i.e. tags, to refine this extraction process and get a more valuable term list. - Identification of competences through the automated assignment of the extracted termsto suitable ontology concepts. - Quantify the competences by means of a novel semantic matching algorithm utilizing the concepts and relations of the ontology. The nodes and edges in the ontology graph are weighted. With the aggregation of these values based on heuristic rules, we generate a user¿s competence profile, which consists of the identified competences, its quantification values and a corresponding reliability indicator. - Develop a user interface in order to enter contributions to the social software system. Contributions may have different forms (e.g., challenges, solutions). Users can easily contribute their experience and discuss it with others. Furthermore, users are enabled to query their competence profile and give feedback to the calculated competence result. According to a participant at the Dagstuhl seminar on'Semantic Web reflections and future directions'2009, researchers agreed on the importance of user interfaces to access semantic information.We prove the validity of the SeCoMine competence measure by means of an experiment conducted with students in a university environment. We validate the measured competences by comparison with both self-evaluation and supervisor assessment.