UCL  IRIS
Institutional Research Information Service
UCL Logo
Please report any queries concerning the funding data grouped in the sections named "Externally Awarded" or "Internally Disbursed" (shown on the profile page) to your Research Finance Administrator. Your can find your Research Finance Administrator at https://www.ucl.ac.uk/finance/research/rs-contacts.php by entering your department
Please report any queries concerning the student data shown on the profile page to:

Email: portico-services@ucl.ac.uk

Help Desk: http://www.ucl.ac.uk/ras/portico/helpdesk
Publication Detail
Online support vector regression for network travel time prediction
Abstract
This study aims to apply a technique of machine learning that is known as Online Support Vector Regression to the forecast of travel time on the London road network. Online SVR is a variation of a more well-known technique called Support Vector Machine (SVM), with two differences. First, where SVM just does the job of classification of the data, SVR models the time series data such that the historical time data is used as an input (independent variable) for regression to determine output (dependent variable). Second and more importantly, it brings in the concept of incremental learning, which is ideally suited to the problem of updating a traffic forecast model in real time. This means that instead of training the system from the start every time, as happens in batch SVR, we are only required to incrementally add or remove data into the system, and the model automatically adapts to the new conditions. The objective of this study is to demonstrate the usability of online SVR for travel time forecast and show that the accuracy of the results of online SVR is comparable to batch SVR. This is achieved by first carrying out analysis of the available data from the London transport network. The experiments are carried out over 10 links strategically selected from the network to represent the whole road network of the capital. It was important to segregate the data into various peak times to better model the trends in the underlying data. The temporal model was devised based on the time series of each of the links. One of the key aims of this study was to actually apply a technique to forecast future values of the time series, not just validate the results using test data. This is attempted with the help of multi-step ahead algorithm, which uses the predicted values iteratively as input in order to forecast future values
Publication data is maintained in RPS. Visit https://rps.ucl.ac.uk
 More search options
UCL Researchers
Editor
Dept of Civil, Environ &Geomatic Eng
Editor
Dept of Civil, Environ &Geomatic Eng
University College London - Gower Street - London - WC1E 6BT Tel:+44 (0)20 7679 2000

© UCL 1999–2011

Search by