Loading…
Welcome to CTRF’s 60th Annual Conference! Enjoy Ottawa
Monday May 26, 2025 2:50pm - 3:10pm EDT
Accurate traffic data imputation is essential for transportation systems, particularly during social events that disrupt traffic patterns. This study introduces the Event-Aware Missing Data Imputation Network (EMDIN), a novel model designed to predict the impact of different types and sizes of social events by leveraging event-specific features, such as type, location, attendance, timings, and parking availability. EMDIN integrates advanced spatiotemporal learning techniques, including graph convolutional networks and long short-term memory (LSTM) layers, enhanced by a two-stage imputation decoder. The model features a dynamic forgetting mechanism in the LSTM layers that prioritizes recent traffic patterns as events approach, while the two-stage decoder further refines missing value predictions using attention-weighted spatiotemporal features. Applied to a one-year probe vehicle and social event dataset from Hamilton, ON, Canada, EMDIN achieved a Mean Absolute Percentage Error (MAPE) of 6.5-7.5% and a Root Mean Square Error (RMSE) of 6.0-6.5 km/h during events, significantly outperforming benchmark methods. This study underscores the importance of incorporating social event features to improve imputation accuracy and provides a robust framework for addressing dynamic urban traffic challenges caused by events.
Speakers
AA

Ali Ardestani

McMaster University
Ali Ardestani is a doctoral candidate in the Civil Engineering department at McMaster University, specializing in Transportation Engineering. He obtained his master’s degree in civil engineering from TMU, Iran with a focus on artificial intelligence application in transportation... Read More →
Monday May 26, 2025 2:50pm - 3:10pm EDT
Desmarais 3105 55 Laurier Ave E, Ottawa ON K1N 6N5

Sign up or log in to save this to your schedule, view media, leave feedback and see who's attending!

Share Modal

Share this link via

Or copy link