Constructing expert profiles over time for skills management and expert finding

Abstract

Expert profiling and identification are important to both knowledge workers and organizations. To this end, skills management and expert recommender systems facilitate the management of skills and competencies and help find appropriate experts to meet a particular need. In this paper, we present a technique for generating evolving expert profiles of individuals composed of their skills and competencies using heterogeneous data from divergent sources of information. We use selfdeclarations, completed learning activities, and previous work experience to generate the initial profile. Recommendations, "wisdom of the crowd", direct observations of online and offline activities, and content generated by the individual both within the organization and on the Web are then used to enrich the profiles and to assess and validate skills and competencies over time.

Expert finding by the Dempster‐Shafer theory for evidence combination

“…Then these methods use an information retrieval technique to rank the related profiles for a given query. Finally, people associated with the ranked profiles are known to be ranked experts (Zhang et al, ; Balog, Azzopardi, & de Rijke, ; Fazel‐Zarandi & Fox, ). In document‐based methods, the information retrieval techniques are used to rank different collections of various documents.…”

Identifying unreliable sources of skill and competency information

“…In particular, references, letters of recommendation, and evaluations in the form of 360 reviews or performance appraisals are important sources used to assess individuals [4], both in hiring new employees and in consideration of promotions to higher ranks.…”