Constructing multi-labelled decision trees for junction design using the predicted probabilities

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

    1 Citation (Scopus)

    Abstract

    In this paper, we evaluate the use of traditional decision tree algorithms CRT, CHAID and QUEST to determine a decision tree which can be used to predict a set of (Pareto optimal) junction design alternatives (e.g. signal or roundabout) for a given traffic demand pattern and available space. This is a multi-label decision tree problem. Traditional decision tree algorithms can normally not deal with multiple target labels, since they aim to produce trees with single target labels. However, we propose an approach in which we normalise the training data and use the predicted probabilities of the resulting tree, confronted with a threshold value, to determine multiple target labels. This enables us to predict sets of junction design alternatives with traditional algorithms and thus having the advantage of using profoundly proven and widely available methods with a range of modelling options. We evaluate our approach based on its performance concerning tree complexity and predictive accuracy, for which we introduce new set comparison measures. We test our approach with different experimental runs varying the algorithms, parameters and threshold values. The results show that it is possible to determine decision trees which can be used to predict sets of junction design alternatives with 82-90% accuracy.
    Original languageEnglish
    Title of host publicationIEEE ITSC 2017
    Subtitle of host publication20th International Conference on Intelligent Transportation Systems: Mielparque Yokohama, Kanagawa, Japan, October 16-19, 2017
    Place of PublicationPiscataway, NJ
    PublisherIEEE
    Pages1-7
    Number of pages7
    ISBN (Electronic)9781538615263
    ISBN (Print)9781538615256
    DOIs
    Publication statusPublished - Oct 2017
    EventIEEE 20th International Conference on Intelligent Transportation Systems (ITSC) - Mielparque Yokohama, Yokohama, Japan
    Duration: 16 Oct 201719 Oct 2017

    Conference

    ConferenceIEEE 20th International Conference on Intelligent Transportation Systems (ITSC)
    Abbreviated titleIEEE ITS 2017
    CountryJapan
    CityYokohama
    Period16/10/1719/10/17

    Fingerprint

    Decision trees
    Labels
    Cathode ray tubes

    Cite this

    Bezembinder, E. M., Wismans, L. J. J., & Van Berkum, E. C. (2017). Constructing multi-labelled decision trees for junction design using the predicted probabilities. In IEEE ITSC 2017: 20th International Conference on Intelligent Transportation Systems: Mielparque Yokohama, Kanagawa, Japan, October 16-19, 2017 (pp. 1-7). Piscataway, NJ: IEEE. https://doi.org/10.1109/ITSC.2017.8317699
    Bezembinder, Erwin M. ; Wismans, Luc J. J. ; Van Berkum, Eric C. / Constructing multi-labelled decision trees for junction design using the predicted probabilities. IEEE ITSC 2017: 20th International Conference on Intelligent Transportation Systems: Mielparque Yokohama, Kanagawa, Japan, October 16-19, 2017. Piscataway, NJ : IEEE, 2017. pp. 1-7
    @inproceedings{64c05dd7913f4d3785be891d8bc10b0d,
    title = "Constructing multi-labelled decision trees for junction design using the predicted probabilities",
    abstract = "In this paper, we evaluate the use of traditional decision tree algorithms CRT, CHAID and QUEST to determine a decision tree which can be used to predict a set of (Pareto optimal) junction design alternatives (e.g. signal or roundabout) for a given traffic demand pattern and available space. This is a multi-label decision tree problem. Traditional decision tree algorithms can normally not deal with multiple target labels, since they aim to produce trees with single target labels. However, we propose an approach in which we normalise the training data and use the predicted probabilities of the resulting tree, confronted with a threshold value, to determine multiple target labels. This enables us to predict sets of junction design alternatives with traditional algorithms and thus having the advantage of using profoundly proven and widely available methods with a range of modelling options. We evaluate our approach based on its performance concerning tree complexity and predictive accuracy, for which we introduce new set comparison measures. We test our approach with different experimental runs varying the algorithms, parameters and threshold values. The results show that it is possible to determine decision trees which can be used to predict sets of junction design alternatives with 82-90{\%} accuracy.",
    author = "Bezembinder, {Erwin M.} and Wismans, {Luc J. J.} and {Van Berkum}, {Eric C.}",
    year = "2017",
    month = "10",
    doi = "10.1109/ITSC.2017.8317699",
    language = "English",
    isbn = "9781538615256",
    pages = "1--7",
    booktitle = "IEEE ITSC 2017",
    publisher = "IEEE",
    address = "United States",

    }

    Bezembinder, EM, Wismans, LJJ & Van Berkum, EC 2017, Constructing multi-labelled decision trees for junction design using the predicted probabilities. in IEEE ITSC 2017: 20th International Conference on Intelligent Transportation Systems: Mielparque Yokohama, Kanagawa, Japan, October 16-19, 2017. IEEE, Piscataway, NJ, pp. 1-7, IEEE 20th International Conference on Intelligent Transportation Systems (ITSC), Yokohama, Japan, 16/10/17. https://doi.org/10.1109/ITSC.2017.8317699

    Constructing multi-labelled decision trees for junction design using the predicted probabilities. / Bezembinder, Erwin M.; Wismans, Luc J. J.; Van Berkum, Eric C.

    IEEE ITSC 2017: 20th International Conference on Intelligent Transportation Systems: Mielparque Yokohama, Kanagawa, Japan, October 16-19, 2017. Piscataway, NJ : IEEE, 2017. p. 1-7.

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

    TY - GEN

    T1 - Constructing multi-labelled decision trees for junction design using the predicted probabilities

    AU - Bezembinder, Erwin M.

    AU - Wismans, Luc J. J.

    AU - Van Berkum, Eric C.

    PY - 2017/10

    Y1 - 2017/10

    N2 - In this paper, we evaluate the use of traditional decision tree algorithms CRT, CHAID and QUEST to determine a decision tree which can be used to predict a set of (Pareto optimal) junction design alternatives (e.g. signal or roundabout) for a given traffic demand pattern and available space. This is a multi-label decision tree problem. Traditional decision tree algorithms can normally not deal with multiple target labels, since they aim to produce trees with single target labels. However, we propose an approach in which we normalise the training data and use the predicted probabilities of the resulting tree, confronted with a threshold value, to determine multiple target labels. This enables us to predict sets of junction design alternatives with traditional algorithms and thus having the advantage of using profoundly proven and widely available methods with a range of modelling options. We evaluate our approach based on its performance concerning tree complexity and predictive accuracy, for which we introduce new set comparison measures. We test our approach with different experimental runs varying the algorithms, parameters and threshold values. The results show that it is possible to determine decision trees which can be used to predict sets of junction design alternatives with 82-90% accuracy.

    AB - In this paper, we evaluate the use of traditional decision tree algorithms CRT, CHAID and QUEST to determine a decision tree which can be used to predict a set of (Pareto optimal) junction design alternatives (e.g. signal or roundabout) for a given traffic demand pattern and available space. This is a multi-label decision tree problem. Traditional decision tree algorithms can normally not deal with multiple target labels, since they aim to produce trees with single target labels. However, we propose an approach in which we normalise the training data and use the predicted probabilities of the resulting tree, confronted with a threshold value, to determine multiple target labels. This enables us to predict sets of junction design alternatives with traditional algorithms and thus having the advantage of using profoundly proven and widely available methods with a range of modelling options. We evaluate our approach based on its performance concerning tree complexity and predictive accuracy, for which we introduce new set comparison measures. We test our approach with different experimental runs varying the algorithms, parameters and threshold values. The results show that it is possible to determine decision trees which can be used to predict sets of junction design alternatives with 82-90% accuracy.

    U2 - 10.1109/ITSC.2017.8317699

    DO - 10.1109/ITSC.2017.8317699

    M3 - Conference contribution

    SN - 9781538615256

    SP - 1

    EP - 7

    BT - IEEE ITSC 2017

    PB - IEEE

    CY - Piscataway, NJ

    ER -

    Bezembinder EM, Wismans LJJ, Van Berkum EC. Constructing multi-labelled decision trees for junction design using the predicted probabilities. In IEEE ITSC 2017: 20th International Conference on Intelligent Transportation Systems: Mielparque Yokohama, Kanagawa, Japan, October 16-19, 2017. Piscataway, NJ: IEEE. 2017. p. 1-7 https://doi.org/10.1109/ITSC.2017.8317699