Daily traffic volume data are collected and stored as historical data. By learning from the historical data, we can predict traffic volumes. In this paper, we propose a clustering method based on the mixture model estimation approach that was introduced in previous papers. This method is compared with the whole-curve-based clustering method. From the method we propose, we derive a partial clustering approach based on the components of the mixture model which was introduced before. The partial clustering method based on components is interesting for research that only focuses on single component. The comparison between methods shows that the mixture-model-based method can reach the results of 7.38% to 14.57% of relative errors compared with the whole-curve-based method.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Mixture-Model-Based Clustering for Daily Traffic Volumes


    Contributors:
    Hu, Yu (author) / Hellendoorn, Hans (author)


    Publication date :

    2015-09-01


    Size :

    1975676 byte





    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English





    Prediction of Daily Traffic Volumes by Using Autoregressive Models

    Suzuki, Y. / IEEE / Institute of Electrical Engineers of Japan et al. | British Library Conference Proceedings | 1999


    Model-Based Motion Clustering Using Boosted Mixture Modeling

    Pavlovic, V. / IEEE Computer Society | British Library Conference Proceedings | 2004