TRB Pubsindex
Text Size:
x

« Prev Next »

Return to Results | New Search

Bookmark and Share

Title:

Tour-Based Mode Choice Study Through Support Vector Machine Classifiers

Accession Number:

01626412

Record Type:

Component

Abstract:

A new approach in recognizing travel mode choice patterns is proposed in this paper. A classification technique, Support Vector Machine (SVM), is used to analyze a tour-based travel demand dataset derived from the 2009 National Household Travel Survey and reporting tours done during a day in the New York State. Three main categories of travel means are considered, i.e. individual motorized means, public transport, bicycle and walking, together with their combinations. Nine variables, both socioeconomic and travel-related, are defined for each personal tour and are then used to build the features space in which the SVM operated its classification process. Results obtained considering different combinations of both kinds of features demonstrate how SVM is able to predict to some extent, in a real settings where car use dominates, which tours are likely to be made by public transport or non-motorized means. Moreover, the flexibility of the technique allows assessing the predictive power of each feature according to the combination of travel means used in different tours. Potential applications of this innovative approach range from activity-based travel choices simulators to search engines supporting personalized travel planners, in general whenever “best guesses” on mode choice patterns have to be quickly made on large amount of data prejudicing the possibility of setting up a statistical model.

Supplemental Notes:

This paper was sponsored by TRB committee ADB10 Standing Committee on Traveler Behavior and Values.

Monograph Accession #:

01618707

Report/Paper Numbers:

17-01885

Language:

English

Corporate Authors:

Transportation Research Board

500 Fifth Street, NW
Washington, DC 20001 United States

Authors:

Pirra, Miriam
Diana, Marco

Pagination:

15p

Publication Date:

2017

Conference:

Transportation Research Board 96th Annual Meeting

Location: Washington DC, United States
Date: 2017-1-8 to 2017-1-12
Sponsors: Transportation Research Board

Media Type:

Digital/other

Features:

Figures; References; Tables

Geographic Terms:

New York (State)

Subject Areas:

Highways; Pedestrians and Bicyclists; Planning and Forecasting; Public Transportation

Source Data:

Transportation Research Board Annual Meeting 2017 Paper #17-01885

Files:

TRIS, TRB, ATRI

Created Date:

Dec 8 2016 10:40AM