Implementación de Banderas Rojas en la Contratación Pública en Colombia utilizando Datos Abiertos
Cargando...
Fecha
Título de la revista
ISSN de la revista
Título del volumen
Editor
Resumen
Este trabajo de grado examina la contratación pública en Colombia, destacando la importancia de mejorar su transparencia y eficiencia mediante técnicas de aprendizaje automático. El estudio utiliza datos abiertos del SECOP II y algoritmos como Isolation Forest para detectar anomalías en los contratos públicos, enfocándose en tres banderas rojas: NF003, que identifica licitaciones con períodos inusualmente cortos; NF016, que detecta valores de licitación anómalos; y NF018, que señala procesos con una sola oferta. Los resultados indican que el 45.16% de los contratos presentan al menos una bandera roja, siendo NF018 la más frecuente. La metodología aplicada permite generar alertas tempranas para facilitar intervenciones oportunas y fortalecer la supervisión de los recursos públicos.
This graduate work examines public procurement in Colombia, highlighting the importance of improving its transparency and efficiency through machine learning techniques. The study uses open data from SECOP II and algorithms such as Isolation Forest to detect anomalies in public contracts, focusing on three red flags: NF003, which identifies bids with unusually short periods; NF016, which detects anomalous bid values; and NF018, which points out processes with only one bid. The results indicate that 45.16% of the contracts present at least one red flag, with NF018 being the most frequent. The methodology applied allows the generation of early warnings to facilitate timely interventions and strengthen the supervision of public resources.
This graduate work examines public procurement in Colombia, highlighting the importance of improving its transparency and efficiency through machine learning techniques. The study uses open data from SECOP II and algorithms such as Isolation Forest to detect anomalies in public contracts, focusing on three red flags: NF003, which identifies bids with unusually short periods; NF016, which detects anomalous bid values; and NF018, which points out processes with only one bid. The results indicate that 45.16% of the contracts present at least one red flag, with NF018 being the most frequent. The methodology applied allows the generation of early warnings to facilitate timely interventions and strengthen the supervision of public resources.