Clustering and Modeling of Network level Traffic States based on Locality Preservative Non-negative Matrix Factorization

Yufei Han (1) , Fabien Moutarde (1)

1 CAOR - Centre de Robotique

In this paper, we propose to cluster and model network-level traffic states based on a geometrical weighted similarity measure of network-level traffic states and locality preservative non-negative matrix factorization. The geometrical weighted similarity measure makes use of correlation between neighboring roads to describe spatial configurations of global traffic patterns. Based on it, we project original high-dimensional network-level traffic information into a feature space of much less dimensionality through the matrix factorization method. With the obtained low-dimensional representation of global traffic information, we can describe global traffic patterns and the evolution of global traffic states in a flexible way. The experiments prove validity of our method for the case of large-scale traffic network.

Domaines: Machine Learning [stat.ML]

Yufei Han, Fabien Moutarde. Clustering and Modeling of Network level Traffic States based on Locality Preservative Non-negative Matrix Factorization. 8th Intelligent Transport Systems (ITS) European Congress, Jun 2011, Lyon, France. ⟨hal-00638074⟩