Árboles de Decisión para el análisis de los defectos técnicos de los ómnibus Yutong

The accidents of transit, annually charge 1.3 million worldwide lives. The factors that have the greatest impact on the accident rate are: human, infrastructure and technical defects. Technical defects represent 8% of the total number of traffic accidents. In Cuba, in order to reduce accidents due to technical defects, an Automobile Technical Inspection is carried out by means of an automated system that detects any technical defect that constitutes a risk to road circulation, which only allows evaluating one vehicle at a time, not being able to assess in a systemic way the incidence of defects in a group of vehicles, which does not allow making informed decisions. The failure to take advantage of the information provided by this system, from a systemic approach, is the fundamental contradiction, making it impossible to make properly argued decisions based on the information stored in its Database Management System. In the present work, alternatives are sought to solve the contradiction, applying Decision Trees techniques, specifically the CHAID model, through which non-explicit knowledge is extracted about the technical defects that are affecting the vehicle fleet under study. A sample of 222 Yutong buses, model HB 6100 ZK, was taken, obtaining five models that identify the affectation of the vehicle fleet, providing a novel solution that contributes to adequately argue the decision making process. Los accidentes del tránsito, anualmente cobran 1,3 millones de vidas a nivel mundial. Los factores que más inciden en la accidentabilidad son: el humano, la infraestructura y los defectos técnicos. Los defectos técnicos representan un 8% del total de los accidentes de tránsito. En Cuba con el objetivo de disminuir los accidentes por defectos técnicos, se realiza una Inspección Técnica Automotor, mediante un sistema automatizado que detecta cualquier defecto técnico que constituya riesgo a la circulación vial, el cual solo permite evaluar un vehículo cada vez, no pudiéndose valorar de manera sistémica la incidencia de los defectos en un grupo de vehículos, lo que no permite tomar decisiones fundamentadas. El no aprovechamiento de la información que brinda este sistema, desde un enfoque sistémico, es la contradicción fundamental, imposibilitando la toma de decisiones debidamente argumentadas a partir de la información guardada en su Sistema de Gestión de Bases de Datos. En el presente trabajo se buscan alternativas para solucionar la contradicción planteada, se a