Adaptive learning in bayesian networks for incident duration prediction

Sami Demiroluk, Kaan Ozbay

Research output: Contribution to journalArticle

Abstract

The development of a practical model for incident management is investigated through Bayesian networks (BNs) in this study. BNs are capable of accurately predicting incident durations and can easily be incorporated into incident management activities of traffic management centers to improve the real-time decision-making process. Three structure learning algorithms were used to construct BN structures. They were estimated by using 2005 New Jersey incident data; the best-performing one was chosen for the incident duration prediction with the use of the 10-fold cross-validation method and the Bayesian information criterion statistic. To demonstrate the performance of Bayesian learning, the chosen model was fed by 2011 New Jersey incident data on a monthly and quarterly basis. Comparing the prediction results for 2011 data with and without adaptive learning showed that the developed BN had the capability to automatically adapt itself to future conditions by learning the patterns of new incidents and their respective durations.

Original languageEnglish (US)
Pages (from-to)77-85
Number of pages9
JournalTransportation Research Record
Volume2460
Issue number1
DOIs
StatePublished - Dec 1 2014

Fingerprint

Bayesian networks
Learning algorithms
Decision making
Statistics

ASJC Scopus subject areas

  • Civil and Structural Engineering
  • Mechanical Engineering

Cite this

Adaptive learning in bayesian networks for incident duration prediction. / Demiroluk, Sami; Ozbay, Kaan.

In: Transportation Research Record, Vol. 2460, No. 1, 01.12.2014, p. 77-85.

Research output: Contribution to journalArticle

Demiroluk, Sami ; Ozbay, Kaan. / Adaptive learning in bayesian networks for incident duration prediction. In: Transportation Research Record. 2014 ; Vol. 2460, No. 1. pp. 77-85.
@article{c9800dd3bb8747b889a5129dfb7d8535,
title = "Adaptive learning in bayesian networks for incident duration prediction",
abstract = "The development of a practical model for incident management is investigated through Bayesian networks (BNs) in this study. BNs are capable of accurately predicting incident durations and can easily be incorporated into incident management activities of traffic management centers to improve the real-time decision-making process. Three structure learning algorithms were used to construct BN structures. They were estimated by using 2005 New Jersey incident data; the best-performing one was chosen for the incident duration prediction with the use of the 10-fold cross-validation method and the Bayesian information criterion statistic. To demonstrate the performance of Bayesian learning, the chosen model was fed by 2011 New Jersey incident data on a monthly and quarterly basis. Comparing the prediction results for 2011 data with and without adaptive learning showed that the developed BN had the capability to automatically adapt itself to future conditions by learning the patterns of new incidents and their respective durations.",
author = "Sami Demiroluk and Kaan Ozbay",
year = "2014",
month = "12",
day = "1",
doi = "10.3141/2460-09",
language = "English (US)",
volume = "2460",
pages = "77--85",
journal = "Transportation Research Record",
issn = "0361-1981",
publisher = "US National Research Council",
number = "1",

}

TY - JOUR

T1 - Adaptive learning in bayesian networks for incident duration prediction

AU - Demiroluk, Sami

AU - Ozbay, Kaan

PY - 2014/12/1

Y1 - 2014/12/1

N2 - The development of a practical model for incident management is investigated through Bayesian networks (BNs) in this study. BNs are capable of accurately predicting incident durations and can easily be incorporated into incident management activities of traffic management centers to improve the real-time decision-making process. Three structure learning algorithms were used to construct BN structures. They were estimated by using 2005 New Jersey incident data; the best-performing one was chosen for the incident duration prediction with the use of the 10-fold cross-validation method and the Bayesian information criterion statistic. To demonstrate the performance of Bayesian learning, the chosen model was fed by 2011 New Jersey incident data on a monthly and quarterly basis. Comparing the prediction results for 2011 data with and without adaptive learning showed that the developed BN had the capability to automatically adapt itself to future conditions by learning the patterns of new incidents and their respective durations.

AB - The development of a practical model for incident management is investigated through Bayesian networks (BNs) in this study. BNs are capable of accurately predicting incident durations and can easily be incorporated into incident management activities of traffic management centers to improve the real-time decision-making process. Three structure learning algorithms were used to construct BN structures. They were estimated by using 2005 New Jersey incident data; the best-performing one was chosen for the incident duration prediction with the use of the 10-fold cross-validation method and the Bayesian information criterion statistic. To demonstrate the performance of Bayesian learning, the chosen model was fed by 2011 New Jersey incident data on a monthly and quarterly basis. Comparing the prediction results for 2011 data with and without adaptive learning showed that the developed BN had the capability to automatically adapt itself to future conditions by learning the patterns of new incidents and their respective durations.

UR - http://www.scopus.com/inward/record.url?scp=84908369142&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84908369142&partnerID=8YFLogxK

U2 - 10.3141/2460-09

DO - 10.3141/2460-09

M3 - Article

VL - 2460

SP - 77

EP - 85

JO - Transportation Research Record

JF - Transportation Research Record

SN - 0361-1981

IS - 1

ER -