Skip to main navigation Skip to search Skip to main content

An Improved K-Nearest Neighbor Model for Road Speed Forecast Based on Spatiotemporal Correlation

  • Beihang University

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Most k-nearest neighbor models only focus on single step short-Term traffic forecasting andcan't perform well when extended to multi-stepforecast. To enhance accuracy, this paper presents an improved k-nearest neighbor model considering thespatiotemporal correlation. The proposed model defines the current conditionsbythetwo-dimensionalspatiotemporalstate matrices, instead of the one-dimensional state vectorof thetime series. Moreover, this paperdeterminesthe weights by Gaussian functionto adjust the matching distance andmanage the data of the nearest neighbors. The original speed data used in this paper arenormalized in order to apply the proposed model to different types of road segments. The case showsthe improved model performs more desirablethan the original k-nearest neighbormodels and demonstrates more appropriate for multi-step forecast ofthe roadspeed.

Original languageEnglish
Title of host publicationCICTP 2015 - Efficient, Safe, and Green Multimodal Transportation - Proceedings of the 15th COTA International Conference of Transportation Professionals
EditorsXuedong Yan, Yu Zhang, Yafeng Yin
PublisherAmerican Society of Civil Engineers (ASCE)
Pages342-351
Number of pages10
ISBN (Electronic)9780784479292
DOIs
StatePublished - 2015
Event15th COTA International Conference of Transportation Professionals: Efficient, Safe, and Green Multimodal Transportation, CICTP 2015 - Beijing, China
Duration: 24 Jul 201527 Jul 2015

Publication series

NameCICTP 2015 - Efficient, Safe, and Green Multimodal Transportation - Proceedings of the 15th COTA International Conference of Transportation Professionals

Conference

Conference15th COTA International Conference of Transportation Professionals: Efficient, Safe, and Green Multimodal Transportation, CICTP 2015
Country/TerritoryChina
CityBeijing
Period24/07/1527/07/15

Keywords

  • K-nearest neighbor
  • Spatiotemporal correlation
  • Speed forecast
  • Weights

Fingerprint

Dive into the research topics of 'An Improved K-Nearest Neighbor Model for Road Speed Forecast Based on Spatiotemporal Correlation'. Together they form a unique fingerprint.

Cite this