Comparative Analysis of Machine Learning Models in the Task of Car Park Congestion Forecasting

The article considers an applied approach to forecasting short-term occupancy of urban parking zones on the example of Moscow. A static model and two recurrent neural networks are compared to analyse time series of parking occupancy percentage. Parking occupancy data undergoes the stages of resampling, skip interpolation and feature engineering. The models are trained on historical data with separation into training, validation and test samples by time; validation is performed using early stopping. Experiments show that the recurrent models outperform the static model in terms of prediction quality. The practical significance of the work lies in the possibility of integrating the proposed models into an intelligent parking management system, which will reduce the time of searching for parking spaces and optimise the load on urban roads in real time.

1. Shoup D.C. Cruising for Parking / D.C. Shoup // Transport Policy. – 2006. – Vol. 13, Iss. 6. – P. 479–486.

2. TomTom Traffic Index 2016 – The Results Are In! Mexico City Takes Crown of 'Most Traffic Congested City' in World from Istanbul // TomTom [Электронный ресурс]. – URL: https://corporate.tomtom.com/static-files/5289c5aa-310c-4965-a4c0-516760a8a6fd (дата обращения: 04.03.2025).

3. TomTom Traffic Index 2017: Mexico City Retains Crown of 'Most Traffic Congested City' in World // TomTom [Электронный ресурс]. – URL: https://corporate.tomtom.com/static-files/7b98ada1-ec2c-4da0-8579-f83ac0f31d88 (дата обращения: 04.03.2025).

4. TomTom Traffic Index 2018 [Электронный ресурс]. – URL: https://traffic-index-docs.s3-eu-west-1.amazonaws.com/TomTomTrafficIndex-Ranking-2018-full.pdf (дата обращения: 04.03.2025).

5. TomTom Traffic Index 2019 [Электронный ресурс]. – URL: https://nonews.co/wp-content/uploads/2020/02/TomTom2019.pdf (дата обращения: 04.03.2025).

6. TomTom Traffic Index 2020 [Электронный ресурс]. – URL: https://traffic-index-docs.s3-eu-west-1.amazonaws.com/TomTomTrafficIndex-Ranking-2020-full.pdf (дата обращения: 04.03.2025).

7. TomTom Traffic Index 2021 [Электронный ресурс]. – URL: https://nonews.co/wp-content/uploads/2022/02/TomTom2021.pdf (дата обращения: 04.03.2025). 

8. United States Patent Application Publication No. US 2004/0032342 A1. Method and System for Projecting Dynamic Parking Availability Based on an Ongoing Survey for Remote Lots With High Demand: Pub. Date: 19.02.2004 / Dunning A.E.

9. Parking Spaces Demand Forecasting Based on ARIMA Model / H. Chen, M. Yang, Yu. Zhang, H. Wen // ICCMS '23: Proceedings of the 2023 15th International Conference on Computer Modeling and Simulation. – New York: Association for Computing Machinery, 2023. – P. 133–137. 

10. Truck Parking Occupancy Prediction: XGBoost-LSTM Model Fusion / S. Gutmann, Ch. Maget, M. Spangler, K. Bogenberger // Frontiers in Future Transportation. – 2021. – Vol. 2. – URL: https://doi.org/10.3389/ffutr.2021.693708 (дата обращения: 27.03.2025).

11. Short-Term Prediction of On-Street Parking Occupancy Using Multivariate Variable Based on Deep Learning / M. Lyu, Ya. Ji, Ch. Kuai, Sh. Zhang // Journal of Traffic and Transportation Engineering. – 2024. – Vol. 11, Iss. 1. – P. 28–40.

12. Multi-Horizon Air Pollution Forecasting with Deep Neural Networks / M. Arsov, E. Zdravevski, P. Lameski // Sensors. – 2021. – Vol. 21, No. 4. – URL: https://doi.org/10.3390/s21041235 (дата обращения: 27.03.2025).

13. Zhao X. Enhancing Predictive Models for On-Street Parking Occupancy: Integrating Adaptive GCN and GRU with Household Categories and POI Factors / X. Zhao, M. Zhang // Mathematics. – 2024. – Vol. 12, No. 18. – URL: https://doi.org/10.3390/math12182823 (дата обращения: 27.03.2025).

14. Wang Zh. Research on Parking Space Prediction Based on Long Short-Term Memory / Zh. Wang, W. Xue, X. Hei // 2019 International Conference on Intelligent Computing, Automation and Systems (ICICAS). – IEEE, 2019. – P. 392–396.

15. Zaremba W., Sutskever I., Vinyals O. Recurrent Neural Network Regularization // arXiv [Электронный ресурс]. – URL: https://arxiv.org/abs/1409.2329 (дата обращения: 27.03.2025).