Prevendo e identificando as causas da evasão do empregado com técnicas de aprendizado de máquina

Abstract

Employee turnover is considered one of the most common issues handled by companies´ HR department. This scenario impacts companies by both direct cost (e.g. advertising for a job opening and recruitment & selection process) and indirect costs (e.g. learning curve and pressure on remaining staff). Also, the costs persist until the performance of the new employee reaches the same level as the other cooworkers, which affect directlythe efficiency of the company. In this sense, this article aims to predict the employee turnover and identify the main reasons that would cause it. Applying the experimental and statistical research methods, the results obtained with the use of data science techniques it were possible through a predictive model with good generalizability for the worked data set, generating satisfactory answers for the prediction of the target event - employee turnover.