基於負二項模型之短期都市交通流量預測

Abstract

Intelligent Transportation Systems (ITS) and Advance Traveler Information Systems (ATIS) require accurate and efficient traffic flow prediction models to support online, realtime and proactive applications. Prediction accuracy is impacted by a critical characteristic of traffic flow that is overdispersion. Efficiency depends on the time complexity of forecasting algorithms. Therefore, this research proposes novel spatiotemporal multivariate prediction models that are based on the Negative Binomial Models (NBAMs). Negative Binomial is utilized to handle overdispersion. The first proposed model is the Negative Binomial Generalized Linear Model (NBGLM), and the second is the Negative Binomial Additive Model (NBAM). The NBGLM is used to capture the spatial and temporal correlations of traffic flow on multiple correlated roads. The NBAM is used to efficiently smooth nonlinear spatial and temporal variables. To evaluate the models, they are applied to real-world data collected from Taipei city and compared with other prediction models. The results indicate that the proposed models are accurate and efficient approaches in predicting traffic flow in urban context where flow is overdispersed, autocorrelated, and influenced by upstream and downstream roads as well as the daily seasonal patterns that are: low, moderate and high traffic seasons.

Intelligent Transportation Systems (ITS) and Advance Traveler Information Systems (ATIS) require accurate and efficient traffic flow prediction models to support online, realtime and proactive applications. Prediction accuracy is impacted by a critical characteristic of traffic flow that is overdispersion. Efficiency depends on the time complexity of forecasting algorithms. Therefore, this research proposes novel spatiotemporal multivariate prediction models that are based on the Negative Binomial Models (NBAMs). Negative Binomial is utilized to handle overdispersion. The first proposed model is the Negative Binomial Generalized Linear Model (NBGLM), and the second is the Negative Binomial Additive Model (NBAM). The NBGLM is used to capture the spatial and temporal correlations of traffic flow on multiple correlated roads. The NBAM is used to efficiently smooth nonlinear spatial and temporal variables. To evaluate the models, they are applied to real-world data collected from Taipei city and compared with other prediction models. The results indicate that the proposed models are accurate and efficient approaches in predicting traffic flow in urban context where flow is overdispersed, autocorrelated, and influenced by upstream and downstream roads as well as the daily seasonal patterns that are: low, moderate and high traffic seasons.