Title: Analysis of road accident data and determining affecting factors by using regression models and decision trees

Authors: Ali Nazeri; Hanieh GharehGozlu; Farshid Faraji; Shabnam Asakareh

Addresses: Department of Industrial Engineering, Islamic Azad University, Damavand Branch, Damavand, Iran ' Department of Industrial Engineering, Islamic Azad University, Damavand Branch, Damavand, Iran ' Department of Management, Islamic Azad University, Damavand Branch, Damavand, Iran ' Department of Industrial Engineering, Islamic Azad University, Damavand Branch, Damavand, Iran

Abstract: This study analyses the road accident data with the aim to predict the probability of road accidents leading to death and determine the affecting factors. Regression models including logit, probit, complementary log-log, gompertz and decision trees based on the CART algorithm were used to analyse the actual data of the rail road police centre of the country. The results show that the logit regression model is superior to the other models from the perspective of the scales of the health indicator. Also, the variables of day of week, age, shoulder path, road side, road type, road position, maximum speed, belt safety, specific safety equipment, vehicle type and vehicle manufacturer country are among the variables that significantly affect the probability of road deaths, and can be controlled by controlling their levels.

Keywords: road accidents; regression models; decision tree model; accuracy indicator scales.

DOI: 10.1504/IJBIDM.2021.10024451

International Journal of Business Intelligence and Data Mining, 2021 Vol.18 No.4, pp.449 - 471

Received: 22 Apr 2018
Accepted: 03 Nov 2018

Published online: 04 May 2021 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article