Development of Predictive Technology for Damage Caused by Icing on Roads in Gangwon-do

Abstract

The number of traffic and fatal accidents caused by road icing in winter has been increasing every year. Since traffic accidents on frozen sections of road can cause serious human and material damage, comprehensive measures are required. To reduce the damage caused by freezing sections in winter in Gangwon-do, this study developed damage prediction technology using data from past damage cases. Traffic accident data on frost and freezing road conditions from 2007 to 2018 were used, and the technology was developed mainly in Gangwon-do. A logistic regression model was implemented for predicting the presence or absence of damage. Variables were selected through the variable selection method and correlation and comparative analyses were conducted. The model developed using the variable selection method was 0.670, and the correlation was large. The model developed using variables showed 0.659. A preliminary preparation using the results of this study will be helpful in preparing for indirect damage caused by heavy snowfall and in decision-making of the relevant public officials.