All Issue

2025 Vol.43, Issue 3 Preview Page

Article

Study on Expressway Traffic Volume Prediction Based on Socio-Economic Indicators Using Deep Learning

딥러닝을 활용한 사회경제지표 기반 고속도로 교통량 예측 연구

Chulhyun PARK1
,
Junghyun KIM2*
박 철현1
,
김 정현2*
1Master Course, Department of Artificial Intelligence, Sejong University, Seoul 05006, Korea
2Professor, Department of Artificial Intelligence and Data Science, Sejong University, Seoul 05006, Korea
1세종대학교 인공지능학과 석사과정
2세종대학교 인공지능데이터사이언스학과 교수
*Corresponding Author
30 June 2025. pp. 265-281
PDF XML Citation
Abstract

Expressway flow and safety management, crucial to a nation’s economy, are essential components of traffic management. To this end, the Korea Expressway Corporation provides traffic forecasts. This study aims to construct a traffic volume prediction model applicable to all situations by using daily traffic volume data—the most fundamental element for predicting traffic conditions—and incorporating not only commonly used time series and meteorological indicators but also socio-economic indicators. The methodology involves constructing a traffic prediction model using time series-based regression analysis and deep learning models, specifically LSTM and Transformer. The data are classified into two cases based on whether the independent variables are normalized. After predicting traffic volume for each model, the results are compared with actual traffic volume and the Korea Expressway Corporation's forecasted traffic volume. The analysis reveals that the deep learning models, LSTM and Transformer, deliver superior performance, with the Transformer model achieving the highest prediction accuracy in overall performance and in detailed analyses according to day patterns. This is attributed to the incorporation of various socio-economic indicators that reflect changing traffic volumes, as well as the application of deep learning models capable of autonomously learning the characteristics and relationships among variables. These results confirm that accurate traffic volume prediction requires not only the inclusion of socio-economic indicators but also the application of deep learning models.

Keywords
deep learning
LSTM(Long Short-Term Memory)
socio-economic indicators
traffic volume prediction
transformer

국가 경제의 대동맥인 고속도로의 원활한 흐름과 안전 관리는 교통관리를 위한 필수적 요소이다. 한국도로공사는 이를 위해 교통상황을 예측하여 제공하고 있다. 본 연구는 교통상황 예측의 가장 기초가 되는 일교통량 데이터를 사용하여, 일반적으로 사용하는 시계열 및 기상지표 뿐만 아니라, 사회경제지표를 기반으로 모든 상황에 적용 가능한 교통량 예측 모형을 구축하는 것이다. 방법론은 시계열 분석 기반의 회귀분석, 딥러닝 방식인 LSTM, Transformer 모델로 교통량 예측 모형을 구축하였으며, 데이터는 독립변수 정규화 여부를 기반으로 2개 Case로 구분하였다. 각 모델별로 교통량을 예측한 후, 실제 교통량, 한국도로공사의 예측 교통량과 비교하였다. 분석 결과 각 모델별 예측 정확도는 딥러닝 방식인 LSTM, Transformer 모델이 우수한 결과를 나타내었으며, 그 중 Transformer 모델이 전반적 성능과 요일 패턴에 따른 세부 결과 분석에 예측 정확도가 가장 우수하였다. 이는 교통량이 변화하는 다양한 사회지표를 반영하였음과 함께, 변수들의 특징과 관계를 스스로 학습할 수 있는 딥러닝 모델을 적용한 결과로 판단된다. 이러한 결과를 통해 정확한 교통량 예측을 위해서는 사회경제지표를 포함하는 것 뿐만 아니라 딥러닝 모델의 적용이 필요함을 확인할 수 있다.

키워드
딥러닝
LSTM(Long Short-Term Memory)
사회경제지표
교통량 예측
transformer
References
1

Baek J. H., Jang H. H., Yoon B. J. (2020), Non-parametric Approach for Prediction of Traffic Volume, The 83rd Conference of KST, Korean Society of Transportation, 236-237.

2

Baek J. H., Yoon B. J. (2019), A Study on Predictive Analysis of Traffic Volume at Highway Tollgate by ARIMA, 2019 Academic Presentation of the Korean Society of Disaster Information, The Korean Society of Disaster Information, 199-200.

3

Chang H. H., Yoon B. J. (2018), Multiple Period Forecasting of Motorway Traffic Volumes by Using Big Historical Data, KSCE J. Civ. Eng., 38(1), Korean Society of Civil Engineers, 73-80.

4

Cho M. W., Chae H. S., Hwang S. H., Jung H. K. (2023), VDS-based Traffic Volume Prediction System Using LSTM-GRU Model, Jour. of KoCon.a, 23(11), The Korea Contents Association, 197-204.

10.5392/JKCA.2023.23.11.197
5

Das L.C. (2023), Traffic Volume Prediction using Memory-Based Recurrent Neural Networks: A Comparative Analysis of LSTM and GRU, arXiv preprint arXiv:2303.12643.

6

Ding W., Zhang T., Wang Z. (2023), Phased Deep Spatio-temporal Learning for Highway Traffic Volume Prediction, arXiv preprint arXiv:2308.06155.

7

Duraku R., Ramadani R. (2019), Development of Traffic Volume Forecasting Using Multiple Regression Analysis and Artificial Neural Network, C.E.J, 5(8), Salehan Institute of Higher Education, 1698-1713.

10.28991/cej-2019-03091364
8

Han D. C., Lee D. W., Jung D. Y. (2021), A Study on the Traffic Volume Correction and Prediction Using SARIMA Algorithm, J. Korean Soc. Intell. Transp. Syst., 20(6), The Korean Society of Intelligent Transport Systems, 1-13.

10.12815/kits.2021.20.6.1
9

Hochreiter S., Schmidhuber J. (1997), Long Short-Term Memory, Neural Computation, 9(8), MIT Press, 1735-1780.

10.1162/neco.1997.9.8.17359377276
10

Jeong S. Y., Beak S. G. (2012), A Model for Forecasting Annual Trips on Expressway, Transportation Technology and Policy, 9(1), Korean Society of Transportation, 95-100.

11

Lee I. J., Lee Y. K. (2021), Functional Regression Approach to Traffic Analysis, KJAS, 34(5), The Korean Statistical Society, 773-794.

12

Lee S. Y., Jung S. Y. (2018), A Study on the Prediction of Traffic Volume on Highway by the Reference Day of Archived Data, Journal of the Society of Disaster Information, 14(2), The Korean Society of Disaster Information, 230-237.

13

Lee Y. S., Jin H. S., Kim Y. J., Park S. H., Yun I. S. (2023), Methodology for Developing a Predictive Model for Highway Traffic Information Using LSTM, J. Korean Soc. Intell. Transp. Syst., 22(5), The Korean Society of Intelligent Transport Systems, 1-18.

10.12815/kits.2023.22.5.1
14

Lu Z., Lv W., Xie Z., Zhu T. (2018), Highway Traffic Volume Prediction via Stacking KNN, SVR, MLP, RNN, 2018 IEEE SmartWorld, Ubiquitous Intelligence & Computing, Advanced & Trusted Computing, Scalable Computing & Communications, Cloud & Big Data Computing, Internet of People and Smart City Innovations (SmartWorld/SCALCOM/UIC/ATC/CBDCom/IOP/SCI), 1408-1413.

10.1109/SmartWorld.2018.00244PMC5905896
15

Min G. G., Ka D. J., Chung S. Y., Lee C. W. (2023), Development of an Explainable Traffic Flow Prediction Model Using Shapley Additive Explanation and Elliptic Bivariate Relationship, J. Korean Soc. Transp., 41(2), Korean Society of Transportation, 239-253.

10.7470/jkst.2023.41.2.239
16

Montgomery D. C., Runger G. C. (2018), Applied Statistics and Probability for Engineers (7th ed.), John Wiley & Sons. 281.

17

Vaswani A., Shazeer N., Parmar N., Uszkoreit J., Jones L., Gomez A. N. et al. (2017), Attention Is All You Need, NeurIPS, 30, Neural Information Processing Systems Foundation, 5998-6008.

18

Zheng Y., Dong C., Dong D., Wang S. (2021), Traffic Volume Prediction: A Fusion Deep Learning Model Considering Spatial-temporal Correlation, Sustainability, 13(19), MDPI(Multidisciplinary Digital Publishing Institute), 10595.

10.3390/su131910595
Information
  • Publisher :Korean Society of Transportation
  • Publisher(Ko) :대한교통학회
  • Journal Title :Journal of Korean Society of Transportation
  • Journal Title(Ko) :대한교통학회지
  • Volume : 43
  • No :3
  • Pages :265-281
  • Received Date : 2024-11-28
  • Revised Date : 2024-12-07
  • Accepted Date : 2025-03-17
  • DOI :https://doi.org/10.7470/jkst.2025.43.3.265
Journal Informaiton Journal of Korean Society of Transportation Journal of Korean Society of Transportation
Archives
: 2015 Vol.34~
  • NRF
  • KOFST
  • crossref cited-by
  • crosscheck
  • orcid
  • open access
  • ccl
Journal Informaiton Journal Informaiton - close