An application of intelligent noise filtering techniques in demand forecasting for carsharing systems

This study deals with demand forecasting for the carsharing system with multi-station and flexible returning station and time. Neural network is employed as the simulation model to forecast the demand at each station at certain time period in the system. This study introduces intelligent filtering techniques as a tool to remove the noise of the data before it is fed into the simulation model. Two filtering techniques have been tested, namely outlier analysis and cluster analysis. Results show that outlier analysis is better compared to cluster analysis in enhancing the accuracy of forecasting model. This shows that proper choice of techniques is important to guarantee that the introduction of this extra procedure could improve the forecasting accuracy.

• Authors:
  • KHOO, H L
  • FUNG, C H
  • MENG, Q
  • LEE, D -
• Publication Date: 2008

Language

• English

Media Info

• Pagination: 3-10
• Serial:
  • INTERNATIONAL JOURNAL OF ITS RESEARCH
  • Volume: 6
  • Issue Number: 1

Subject/Index Terms

• TRT Terms: Automobiles; Demand; Errors; Forecasting; Intelligent transportation systems; Mathematical models; Motor vehicles; Ownership; Paratransit services; Simulation
• ITRD Terms: 1243: Car; 1157: Community transport; 285: Demand (econ); 1157: Dial a ride; 6440: Error; 132: Forecast; 8735: Intelligent transport system; 6473: Mathematical model; 9103: Simulation; 315: Vehicle ownership
• Subject Areas: Operations and Traffic Management; Public Transportation; I70: Traffic and Transport; I90: Vehicles;

Filing Info

• Accession Number: 01127618
• Record Type: Publication
• Source Agency: TRL
• Files: ITRD
• Created Date: May 4 2009 7:41AM