USO DE TÉCNICAS DE AGRUPAMENTO E REDE NEURAL EM SINISTROS COM MÁQUINAS AGRÍCOLAS NAS RODOVIAS FEDERAIS PARANAENSES

Autores

  • Deivielison Ximenes Siqueira Macedo Universidade Federal do Ceará - UFC
  • Leonardo de Almeida Monteiro Universidade Federal do Ceará - UFC
  • Viviane Castro dos Santos Universidade Federal do Ceará - UFC
  • Enio Costa Universidade Federal do Ceará - UFC
  • Daniel Albiero Universidade Federal do Ceará - UFC

DOI:

https://doi.org/10.17224/EnergAgric.2018v33n1p14-21

Resumo

Acidentes com máquinas agrícolas estão cada vez mais presentes na realidade do trabalhador rural, isso pela grande difusão e necessidade do uso dessas máquinas no campo, todavia muitos desses acidentes são negligenciados na sua forma de ocorrência, havendo necessidade de estudos sobre os indicadores desses sinistros com intuito de tentar evitá-los. O uso associado de técnicas de agrupamento de dados e redes neurais vem se mostrando de grande eficiência para demonstrar o impacto de variáveis sobre a dinâmica do acidente, pois levam em consideração todas as variáveis em conjunto. Assim objetivou-se realizar o uso de redes neurais, análise de agrupamento hierárquico e análise de agrupamento não hierárquico sobre os indicadores dos acidentes envolvendo tratores ocorridos em rodovias federais do estado do Paraná. O trabalho consiste de uma parceria do Laboratório de Investigação de Acidentes com Máquinas Agrícolas-LIMA com 16ª Superintendência da Polícia Rodoviária Federal, o qual forneceu o Boletim de Acidente de Trânsito-BAT referente a máquinas agrícolas nas rodovias federais paranaenses no período de 2007 a 2016. No BAT os indicadores avaliados foram: tipo de acidente, causa do acidente, período do dia em que ocorreu o acidente, condições climáticas no momento do acidente e situação física da vítima. O uso das redes neurais foi realizado através dos mapas auto organizados, os métodos de agrupamento hierárquico foi realizado através do dendrograma e o não hierárquico pelo coeficiente de K-means. A técnica em conjunto permitiu analisar todas as variáveis em conjunto fornecendo uma visão mais clara do processo de ocorrência dos sinistros.

PALAVRAS-CHAVE: Acidente, Estradas, Segurança.

 

USE OF CLUSTER ANALYSIS AND NEURAL NETWORK IN ACCIDENTS WITH AGRICULTURAL MACHINES IN THE FEDERAL ROADS OF PARANÁ

ABSTRACT: Accidents with agricultural machinery are increasingly present in the rural reality, due to the great diffusion and necessity  of these machines use in the field.  However, many of these accidents are neglected in their form of occurrence, and studies on accidents  indicators are necessary to attempt  to avoid them. The associated use of data grouping techniques and neural networks has been shown to be very efficient in demonstrating the impact of variables on the dynamics of  accident, since they take into account all the variables together. The objective was to perform the use of neural networks, hierarchical grouping analysis and non-hierarchical grouping analysis on the indicators of tractors accidents occurred on federal highways in the state of Paraná. The study  consisted  of a partnership between g the Laboratory of Investigation of  Agricultural Machinery Accidents -LIMA,  and the 16th Superintendence of the Federal Highway Police, which provided the Traffic Accident Bulletin-BAT of  agricultural machinery on the federal highways of Paraná from 2007 to 2016. The indicators evaluated were: type of accident, cause of the accident, period of the day in which the accident occurred, climatic conditions at the time of the accident and the physical situation of the victim. The use of neural networks was performed through self-organized maps; hierarchical grouping methods were performed through the dendrogram and non-hierarchical using the K-means coefficient. The technique together allowed analyzing all the variables together providing a clearer view of the process of claims occurrence.

KEYWORDS: Accident, Road, Security

Downloads

Publicado

30-05-2018

Como Citar

Macedo, D. X. S., Monteiro, L. de A., Santos, V. C. dos, Costa, E., & Albiero, D. (2018). USO DE TÉCNICAS DE AGRUPAMENTO E REDE NEURAL EM SINISTROS COM MÁQUINAS AGRÍCOLAS NAS RODOVIAS FEDERAIS PARANAENSES. ENERGIA NA AGRICULTURA, 33(1), 14–21. https://doi.org/10.17224/EnergAgric.2018v33n1p14-21

Edição

Seção

Automação e Otimização de Máquinas e Equipamentos Agrícolas