Prediction and understanding of employee retention: a machine learning application

• Autores: Musso, Mariel Fernanda; Cascallar, Eduardo
• Año de publicación: 2020
• Idioma: inglés
• Tipo de recurso: documento de conferencia
• Materia: Machine learning; Employee retention; Neural networks; Workplace
• Institución: Consejo Nacional de Investigaciones Científicas y Técnicas

The main objectives of this study were to develop accurate predictive models of “employee retention” and to understand the contribution of specific personal and organizational factors predicting this phenomenon. The participants were 993 employees (54.2% female) from different organizations in the private and public sector, age mean: 32 years old (SD= 10.33); seniority: 5.83 years (SD= 6.7). A socio-demographic questionnaire to collect personal background factors and an employee retention questionnaire were applied. Multilayer perceptron artificial neural networks (ANN) with a backpropagation algorithm were developed in order to identify employees with low intention to stay in the current organization (low 33%). ANN achieved a high accuracy in the training testing phase (77%), testing phase (100%), and validation set (100%) for the target group. A more accurate identification of those workers who have a low sense of belonging within the company, would allow a more targeted investment in personnel training.
Fil: Musso, Mariel Fernanda. Universidad Argentina de la Empresa; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas. Oficina de Coordinación Administrativa Saavedra 15. Centro Interdisciplinario de Investigaciones en Psicología Matemática y Experimental Dr. Horacio J. A. Rimoldi; Argentina
Fil: Cascallar, Eduardo. Katholikie Universiteit Leuven; Bélgica
Earli SIG14 2020 Conference
Barcelona
España
European Association for Research of Learning and Instruction
Universitat Autónoma de Barcelona