A predictive model for employee attrition

• Autor(a) principal: Gomes, Adriano de Oliveira
• Data de Publicação: 2022
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10773/35942
• topic: Employee attrition; Machine learning; Real data; Human resources management; Industry 4.0

Employee attrition is currently a major concern for companies as loosing highly qualified personnel has tremendous impacts in several aspects of its daily life. Being able to predict or anticipate attritors is highly valuable in the era of industry 4.0, as it can avoid needless costs. This dissertation proposes an approach to build a database and a predictive model for attrition that includes data from multiple companies from different sectors. Several standards and data definitions are proposed to ease collection and data fusion from different sources, resulting in a database on which a predictive model can be trained. This dissertation’s proposal deals with attrition by considering 3 classes (voluntary, involuntary and no attritors) and several machine learning models were tested to solve the problem. It was found that boosting models stand out as the best performing ones. A XGBoost model evaluated on a 20-run experiment achieved an overall mean accuracy of 78.5%, corresponding to 52.6% of the voluntary attritors, 78.9% of the involuntary attritors and 81.6% of the non-attritors, showing that voluntary attritors are harder to discriminate. For these results, the contract type, area of work in the company or salary rate have shown to be the most important factors contributing to attrition.