The invention discloses a traffic flow adaptive prediction method based on machine learning and collaborative filtering. According to the method, urban road flow is predicted according to entrance lanes and time periods by using a data-driven non-parametric machine learning model library. A global optimal prediction model scheme is abandoned, the optimal model is selected from multiple models, and it is guaranteed that each entrance lane can have the optimal prediction effect in different time periods; and a deep learning model LSTMs with better prediction performance is added into a model library, so that the upper limit of prediction precision is improved. According to the method, the thought of collaborative filtering is utilized, and the effect that all the entrance lanes can obtain the optimal model can be achieved only by using part of the entrance lanes for full-model training. The deep learning model training of all entrance lanes is avoided, and the computing power and the training time are saved.
本发明公开了一种基于机器学习与协同过滤的交通流自适应预测方法。本发明利用数据驱动的非参数机器学习模型库对城市道路流量分进口道、分时段进行预测。摒弃了全局最优预测模型方案,在多种模型中择优进行选择,保证了每个进口道不同时段能有最优的预测效果;并且在模型库中加入了预测性能更好的深度学习模型LSTMs,提高了预测精度上限。本发明利用协同过滤的思想,只须用部分进口道进行全模型训练,即可达到所有进口道获得最优模型的效果。避免了所有进口道进行深度学习模型的训练,节省了算力和训练时间。
Traffic flow adaptive prediction method based on machine learning and collaborative filtering
基于机器学习与协同过滤的交通流自适应预测方法
2023-06-27
Patent
Electronic Resource
Chinese
IPC: | G08G Anlagen zur Steuerung, Regelung oder Überwachung des Verkehrs , TRAFFIC CONTROL SYSTEMS / G06F ELECTRIC DIGITAL DATA PROCESSING , Elektrische digitale Datenverarbeitung / G06N COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS , Rechnersysteme, basierend auf spezifischen Rechenmodellen |
Adaptive traffic light prediction via Kalman filtering
IEEE | 2014
|ADAPTIVE TRAFFIC LIGHT PREDICTION VIA KALMAN FILTERING
British Library Conference Proceedings | 2014
|Adaptive learning traffic flow prediction method under dynamic traffic condition
European Patent Office | 2023
|Adaptive traffic signal control method combining traffic flow prediction and reinforcement learning
European Patent Office | 2023
|