File(s) under permanent embargo
Predicting driving direction with weighted markov model
conference contribution
posted on 2012-12-01, 00:00 authored by B Mao, J Cao, Z Wu, Guangyan HuangGuangyan Huang, J LiDriving direction prediction can be useful in different applications such as driver warning and route recommendation. In this paper, a framework is proposed to predict the driving direction based on weighted Markov model. First the city POI (Point of Interesting) map is generated from trajectory data using weighted PageRank algorithm. Then, a weighted Markov model is trained for the near term driving direction prediction based on the POI map and historical trajectories. The experimental results on real-world data set indicate that the proposed method can improve the original Markov prediction model by 10% at some circumstances and 5% overall. © Springer-Verlag 2012.
History
Volume
7713 LNAIPagination
407 - 418Publisher DOI
ISSN
0302-9743eISSN
1611-3349ISBN-13
9783642355264Publication classification
E Conference publication; E1.1 Full written paper - refereedCopyright notice
2013, SpringerTitle of proceedings
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)Usage metrics
Categories
No categories selectedLicence
Exports
RefWorks
BibTeX
Ref. manager
Endnote
DataCite
NLM
DC