Exploring expressways traffic volume determinants and Prediction with deep learning models

This study aims to explore the main determinants affecting traffic volume on Korean expressways and implement a traffic forecasting model based on them. Through Exploratory Data Analysis of traffic volume data provided by Korea Expressway Corporation, we identified the correlation of major variables and analysed environmental and social factors affecting traffic volume. Based on this, we built deep learning time series prediction models that have advantages in predicting dependent variables with nonlinearity, and compared the performance of each model. The results of the study showed that factors such as precipitation and snowfall in a specific region, month, day of the week, national holidays, and public holidays have a significant impact on highway traffic volume, and Block-rNN, N-HiTS of the deep learning-based models were found to be superior in terms of prediction performance. This study is expected to contribute to the development of efficient policies for highway traffic management and traffic information prediction models.