A Data-driven Method for Competency Evaluation of Personnel.

Nowadays, human capital is very important for organizations. The organizations with more outstanding talents can often gain a foothold in the fierce competition. As a tool to measure talents' performance, competency model often plays a vital role in identifying and evaluating talents. Traditionally, the process of building competency model needs to invest a lot of manpower and time. Also it often lacks of objectivity to some extent. Meanwhile, the original data in the human resource (HR) database of organizations is often not fully utilized. This paper proposes a data-driven method to build competency model, in order to convert original data in database to more valuable information for evaluating talents. Firstly, a data preprocessing framework is designed to facilitate the use of HR data in subsequent analysis. Then 9 methods are designed to construct a set of features that can objectively reflect the situation and abilities of personnel. Data analytics and machine learning are mainly used to construct and verify the competency model. A case study is also included in this paper, a competency model of the middle-level manager (MLM) of a Chinese state-owned enterprise is obtained based on the proposed method. This competency model is also verified by the validation mechanism designed in this paper.