Modelo de perfilación y selección de personal aplicado al sector BPO en Colombia basado en técnicas de Predictive Analytics

Abstract

This paper presents a personnel profiling and selection model using data analysis (DA) tools with a predictive analytics (PA) approach. This approach, based on historical data, sought to determine the appropriate profile for hiring personnel, as well as their optimal assignment to roles. This paper presents a study for the development of a personnel profiling and selection model using computational intelligence techniques to improve the personnel selection process in companies in the Business Process Outsourcing (BPO) sector in Colombia, using DA tools with a PA approach, addressing the critical problem of early turnover. Through a mixed methodology based on qualitative and quantitative methods, key attributes such as age, number of dependents, and work experience were identified, which significantly impact employee retention. Machine learning algorithms were used, highlighting the support vector classification (SVC) algorithm, and techniques such as synthetic minority oversampling (SMOTE) to balance unbalanced classes in the dataset. The developed predictive model reduced early churn from an initial 62.82% to 4.11% during training and to 15.95% in the following month of implementation, equivalent to a 93.60% decrease in this indicator. This result translated into savings of more than $1 billion COP in costs associated with selection, training, and learning curves in one quarter and $750 million COP in the second month. SVC proved to be the most effective algorithm, with a specificity of 94.44% and a minimal incidence of false positives, reinforcing its usefulness in predicting and selecting suitable candidates. The results obtained validate the impact of predictive models on cost reduction and process improvement, and establish a basis for adapting this methodology to other industrial sectors. The study concludes that the integration of predictive techniques can transform human resource (HR) management, offering scalable and adaptive solutions for dynamic work environments.