2020/05/08 by Eric L. Manibardo, Manibardo, Eric L., Ibai Laña +3
Engineering · #68T05 #Artificial Intelligence (cs.AI) #FOS: Computer and information sciences #Machine Learning (cs.LG) #Traffic Prediction and Management Techniques #Traffic control and management #Vehicle emissions and performance
paper · pdf · doi:10.48550/arxiv.2005.05069
openalex publication_date 2020/05/08 · openalex created_date 2022/07/26 · openalex updated_date 2026/07/28
This work aims at unveiling the potential of Transfer Learning (TL) for\ndeveloping a traffic flow forecasting model in scenarios of absent data.\nKnowledge transfer from high-quality predictive models becomes feasible under\nthe TL paradigm, enabling the generation of new proper models with few data. In\norder to explore this capability, we identify three different levels of data\nabsent scenarios, where TL techniques are applied among Deep Learning (DL)\nmethods for traffic forecasting. Then, traditional batch learning is compared\nagainst TL based models using real traffic flow data, collected by deployed\nloops managed by the City Council of Madrid (Spain). In addition, we apply\nOnline Learning (OL) techniques, where model receives an update after each\nprediction, in order to adapt to traffic flow trend changes and incrementally\nlearn from new incoming traffic data. The obtained experimental results shed\nlight on the advantages of transfer and online learning for traffic flow\nforecasting, and draw practical insights on their interplay with the amount of\navailable training data at the location of interest.\n