Uso de redes neuronales para determinar la influencia del estado del pavimento en siniestros viales de la ciudad de Bogotá

Road accidents must be seen as a public health problem. Consequently, any effort to reduce the number of victims is fundamental. This research analyzes the influence of the state of the road network (diagnosed in 2016 by the IDU of the city of Bogotá) in the type of road accidents reported since 2007 by the traffic police to the Secretary of Mobility of the city. This analysis is carried out from the construction of a fully connected multilayer neural network. The input data were obtained from a database that included the location of road accidents and the characteristics of the condition of the pavement. The output is the prediction of the probability of an accident to happen and its severity, classified as only damage, injuries, or deaths. After a sensitivity analysis on the variables associated with the condition of the pavement, it was found that PCI and pavement surface...

• Autores: Ferro Briceño, Paula Vanessa
• Fecha de publicación: 2021
• Institución: Universidad de los Andes
• Repositorio: Séneca: repositorio Uniandes
• Idioma: spa
• OAI Identifier: oai:repositorio.uniandes.edu.co:1992/50973
• Acceso en línea: http://hdl.handle.net/1992/50973
• Palabra clave: Accidentes de tránsito; Infraestructura vial; Pavimentos; Redes neurales (Computadores); Ingeniería
• Rights: openAccess
• License: https://repositorio.uniandes.edu.co/static/pdf/aceptacion_uso_es.pdf