Expertise assessment with multi-cue semantic information

Assessing and managing the expertise of employees in knowledge and service industries is critical because human capital is the key differentiator among companies. Moreover, professional social networks are becoming increasingly popular. Besides the well-known public professional social network site Linked In, enterprise social networks are also now being widely used inside corporations and companies. In this paper, we address the critical workforce analytics problem of automatically assessing employees' skills by mining multiple cues found in enterprise and social data. In particular, we treat the assessment of employees' expertise as a matrix completion problem, where the rows represent individual employees and the columns represent individual skills. The multiple cues about employee expertise come from data we integrate on the existing skill assessment process within the company, the social networking and social media activity of the employees, and the semantic similarity of skills. Assessment results are evaluated as a binary classification recommendation. Extensive empirical study using a real-world data set from a large multinational Fortune 500 corporation corroborates the effectiveness of multi-cue analytics to improve the coverage and accuracy of skill assessment.