Optimization of Hybrid Neural Network Predictor for Spatial Traffic Flow Data Treatment: Genetic Phase-Space Reconstruction
Eleni I. Vlahogianni, Matthew G. Karlaftis · 2004
The present paper proposes a hybrid neural predictor that consists of temporal structures of Multilayer Perceptrons for encompassing traffic flow series from sequential points in a traffic network to improve short-term traffic flow prediction. Each of the temporal structures is genetically optimized to provide the optimum embedding of the series attached.