Events

Prof. Bin KE
Professor | Provost’s Chair
Accounting Department, Business School
National University of Singapore

Due to the rise of machine learning and big data, people analytics is expected to be widely adopted by firms around the world and has the potential to disrupt the human capital management profession. However, due to the proprietary nature of human capital management data, there is little empirical research on the benefits and costs of adopting people analytics in practice. We contribute to this new literature by examining the feasibility of using big data and machine learning to enhance a firm’s human capital management. We leverage the proprietary data from a Chinese firm on static employee characteristics and employee behavioral data to build a machine learning model to predict voluntary employee turnover on a yearly basis. We show that it is possible to combine static employee characteristics with employee behavioral data to build an implementable machine learning model of employee turnover. We find that complicated machine learning methods may not always dominate simple and easy-to-interpret models in the corporate setting with small data sets. There is also evidence that common machine learning models may not work well in real world due to concept drift (e.g., the Covid-19 pandemic) and data drift (e.g., due to change in data collection methods).