Information exploitation of human resource data with persistent homology

Abstract

In an era of big data, corporations have access to an abundance of employee details. While few inferences about employee performance can be made from these data, discarding them may be potentially detrimental to a business. Likewise, employee applications contain substantial amounts of information that cannot necessarily be used to indicate the potential performance of employees should they be appointed. “Persistent homology” considers the topography of data, identifying clusters of behavior that may be associated with performance levels, as well as “holes” in the data cloud that may be filled with suitable job applicants. Therefore, this study presents a theoretical application of persistent homology to human resource management, which considers the topography of data to identify clusters of behavior associated with performance levels and fill gaps with suitable job applicants. Our study demonstrates the potential of persistent homology that offers a breakthrough contribution to the wider research agenda.