A data-driven decision-making framework for personnel selection based on LGBWM and IFNs

For any organization, personnel selection is regarded as one of the most critical and complex issues in human resource management. Due to the fuzziness and ambiguity in personnel selection, it was usually tackled by utilizing multi-criteria decision-making (MCDM) methods. Previously developed MCDM methods for personnel selection have principally depended on experts' preferences, which may lead to the bias and deviation caused by human cognitive limitations. With the development of information technology, many human resource (HR) data have accumulated within some organizations, making it possible to discover meaningful patterns in HR data based on data analytics algorithms (DAAs). These patterns do not depend on experts' judgement and can objectively reflect the performance of personnel in several aspects. Thus, it is promising that integrating objective patterns obtained from HR data and experts' judgement can make personnel selection more systematic and scientific. To tackle this complex scenario, we propose a data-driven decision-making framework that combines DAAs and MCDM method. The proposed framework can explore the underlying patterns in HR data and also consider experts' judgement during the selection process. In this framework, a data-driven competency-based method is designed to mine out valuable information in HR data. In addition, for assisting multiple experts, a linear group best-worst method (LGBWM) is proposed to calculate the weights of criteria, and intuitionistic fuzzy numbers (IFNs) are used to help experts express their preferences. A personnel evaluation and selection (PLEAS) decision support system is also designed and developed to implement the framework better. In order to illustrate the effectiveness of the proposed framework, a real-world case is conducted in a Chinese state-owned company, and the results demonstrate that our proposed framework is valid for solving data-driven personnel selection problems. (C) 2022 Elsevier B.V. All rights reserved.