Drivers' Behavior and Traffic Accident Analysis Using Decision Tree Method

Pires Abdullah

doi:10.3390/su141811339

2022-09-09

Pires Abdullah *

Tibor Sipos

DOI：10.3390/su141811339

原文链接

原文求助

摘要

En 中文

This study was carried out to examine the severity level of crashes by analyzing traffic accidents. The study's goal is to identify the major contributing factors to traffic accidents in connection to driver behavior and socioeconomic characteristics. In order to find the most probable causes in accordance with the major target variable, which is the level of severity of the crash, the study set out to identify the main attributes induced by the decision tree method (DT). The local people received a semi-structured questionnaire interview with closed-ended questions. The survey asked questions about drivers' attitude and behavior, as well as other contributing factors such as time of accidents and road type. The attributes were analyzed using the machine-learning method using DT with Python programming language. This method was able to determine the relationship between severe and non-severe crashes and other significant influencing elements. The Duhok city people participated in the survey, which was conducted in the Kurdistan area of northern Iraq. The results of the study demonstrate that the number of lanes, time of the accident, and human attitudes, represented by their adherence to the speed limit, are the primary causes of accidents with victims.

Keyword:

machine learning

decision tree classifier

severe accident analysis

social factors

road safety