Prédiction de l'état du trafic routier basée sur les motifs et les chaînes de Markov

Feda Almuhisen 1 Nicolas Durand 1 Leonardo Brenner 2 Quafafou Mohamed 1
1 DANA - Data Mining at scale
LIS - Laboratoire d'Informatique et Systèmes
2 MOFED - Modèles et Formalismes à Evénements Discrets
LIS - Laboratoire d'Informatique et Systèmes
Abstract : This paper proposes a new method for predicting traffic state within short time windows. This method takes advantage from space-partitioning, pattern extraction and Markov modelling. From trajectories, frequent regions are extracted where vehicles repeatedly pass through by using the frequent closed patterns and the traffic state is detected based on the evolution of these patterns over time. The next state of traffic for the frequent regions is then predicted based on the Markov models. Experiments on real-world data show that the proposed method is more accurate than a baseline method.
Document type :
Conference papers
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-01858562
Contributor : Nicolas Durand <>
Submitted on : Wednesday, December 12, 2018 - 2:27:52 PM
Last modification on : Tuesday, April 2, 2019 - 9:43:19 AM
Document(s) archivé(s) le : Wednesday, March 13, 2019 - 2:11:02 PM

File

almuhisen_sfc2018.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01858562, version 1

Collections

Citation

Feda Almuhisen, Nicolas Durand, Leonardo Brenner, Quafafou Mohamed. Prédiction de l'état du trafic routier basée sur les motifs et les chaînes de Markov. 25èmes Rencontres de la Société Francophone de Classification (SFC 2018), Sep 2018, Paris, France. ⟨hal-01858562⟩

Share

Metrics

Record views

91

Files downloads

60