Layered ensemble model for short-term traffic flow forecasting with outlier detection

摘要: Real time traffic flow forecasting is a necessary requirement for management in order to be able evaluate the effects of different available strategies or policies. This paper focuses on short-term by taking into consideration both spatial (road links) and temporal (lag past values) information. We propose Layered Ensemble Model (LEM) which combines Artificial Neural Networks Graded Possibilistic Clustering obtaining an accurate forecast rates with outlier detection. Experimentation has been carried out two data sets. The former was obtained from real UK motorway later simulated street network Genoa (Italy). proposed LEM model provides promising results given ability detection, accuracy, robustness approach, it can fruitful integrated systems.