Predicción de la rotación de personal en la empresa Sitel Costa Rica
Loading...
Date
Authors
Ramírez Rodríguez, Sergio Vinicio
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Resumen
El entorno empresarial actual, caracterizado por una competencia intensa en mercados globalizados y la necesidad imperativa de innovación para garantizar
una ventaja competitiva, ha resaltado la importancia del análisis de datos como una herramienta fundamental para optimizar recursos, aumentar las ganancias
y mejorar los indicadores de desempeño en las empresas. Un indicador crucial para cualquier empresa es la minimización de la rotación de empleados, especialmente en sectores como el de Business Process Outsourcing (BPO).
Una alta rotación conlleva un aumento significativo en los costos de reclutamiento, el tiempo requerido para encontrar reemplazos, así como los gastos
asociados con la capacitación de nuevos empleados y la pérdida del conocimiento acumulado por el personal que se marcha, entre otros efectos negativos. Por
lo tanto, es esencial que las empresas mitiguen este riesgo mediante el uso de análisis de datos y metodologías de machine learning para desarrollar modelos
predictivos que identifiquen la probabilidad de rotación de empleados dentro de un período específico.
Se evaluaron diversos modelos, incluidos XGBoost, Bosques Aleatorios, Regresión Logística y Consensus, comparando métricas como el área bajo la curva
ROC, la precisión global, la sensibilidad y la especificidad. XGBoost demostró ser superior debido a su alta capacidad predictiva, aprovechando un enfoque de
ensamblado con arboles de decisión, técnicas de regularización para prevenir el sobreajuste, optimización de hiperparámetros para una configuración óptima
y escalabilidad para manejar conjuntos de datos grandes y complejos. Además, se evaluó la sensibilidad del modelo mediante pruebas de estrés tanto en las
observaciones como en las variables predictoras. Desde su implementación, el modelo ha generado ahorros de millones de dólares en los aspectos mencionados.
Abstract The current business environment, characterized by intense competition in globalized markets and the imperative need for innovation to ensure a competitive advantage, has underscored the importance of data analytics as a fundamental tool for optimizing resources, increasing profits, and enhancing performance indicators in companies. One crucial indicator for any company is the minimization of employee attrition, especially in sectors like Business Process Outsourcing (BPOs). High attrition leads to significant increases in recruitment costs, time required to find replacements, as well as expenses associated with training new employees and the loss of accumulated knowledge by departing staff, among other negative effects. Therefore, it is essential for companies to mitigate this risk by using data analytics and machine learning tools to develop predictive models to identify the likelihood of employee attrition within a specific period. Various models, including XGBoost, Random Forest, Logistic Regression, and Consensus, were evaluated, comparing metrics such as area under the ROC curve, overall accuracy, sensitivity, and specificity. XGBoost emerged as superior due to its adept predictive capacity, leveraging an ensemble approach with decision trees, regularization techniques to forestall overfitting, hyperparameter optimization for optimal configuration, and scalability for handling large and complex datasets. Additionally, the model’s sensitivity was assessed through stress tests on both observations and predictor variables. Since implementation, the model has yielded millions of dollars in the aforementioned savings.
Abstract The current business environment, characterized by intense competition in globalized markets and the imperative need for innovation to ensure a competitive advantage, has underscored the importance of data analytics as a fundamental tool for optimizing resources, increasing profits, and enhancing performance indicators in companies. One crucial indicator for any company is the minimization of employee attrition, especially in sectors like Business Process Outsourcing (BPOs). High attrition leads to significant increases in recruitment costs, time required to find replacements, as well as expenses associated with training new employees and the loss of accumulated knowledge by departing staff, among other negative effects. Therefore, it is essential for companies to mitigate this risk by using data analytics and machine learning tools to develop predictive models to identify the likelihood of employee attrition within a specific period. Various models, including XGBoost, Random Forest, Logistic Regression, and Consensus, were evaluated, comparing metrics such as area under the ROC curve, overall accuracy, sensitivity, and specificity. XGBoost emerged as superior due to its adept predictive capacity, leveraging an ensemble approach with decision trees, regularization techniques to forestall overfitting, hyperparameter optimization for optimal configuration, and scalability for handling large and complex datasets. Additionally, the model’s sensitivity was assessed through stress tests on both observations and predictor variables. Since implementation, the model has yielded millions of dollars in the aforementioned savings.
Description
Keywords
Rotación de personal, Machine Learning, XGBoost, Bosques Aleatorios, Regresión Logística
Citation
Collections
Endorsement
Review
Supplemented By
Referenced By
Creative Commons license
Except where otherwised noted, this item's license is described as Attribution-NonCommercial-NoDerivatives 4.0 International