Due to the rapid update of mobile applications, the characteristics of network traffic may change greatly, which leads to the degradation of network traffic classification model. The network traffic classification model requires accurate labelled data sets. Therefore, it is necessary to build high quality network traffic ground truth for mobile applications. This paper suggests an automated ground truth collection system for mobile application traffic, called MAT-GT, which can collect the network traffic of mobile applications and quickly construct accurate ground truth. It contains four components: controller engine, capture engine, traffic process engine and noise filter engine. MAT-GT mainly uses the user identifier of application and netfilter log to collect and label traffic of the mobile applications. For the labelling of fine-grained behavior traffic, we propose to use the injection timing behavior labelling method to slice the traffic of the application. After the traffic traces processed, we combine the noise recognition model to filter out the noisy traffic in the dataset. Then, we accurately form labelled traffic data sets for 150 mobile applications and 33 behaviors traffic. Finally, we apply non-noisy application traffic datasets for mobile application traffic classification, which further verifies that the accurately labelled data set provided by our system can improve the classification effect of application traffic.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    MAT-GT: An Automated Ground Truth Collection Tool for Mobile Application Traffic


    Beteiligte:
    Yang, Qingya (Autor:in) / Zhang, Jialin (Autor:in) / Shi, Junzheng (Autor:in) / Fu, Peipei (Autor:in) / Xiong, Gang (Autor:in)


    Erscheinungsdatum :

    11.12.2022


    Format / Umfang :

    2494413 byte




    Medientyp :

    Aufsatz (Konferenz)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    Relating Ground Truth Collection to Model Sensitivity

    Amar, F. / Fung, A. K. / Karam, M. A. et al. | British Library Conference Proceedings | 1993



    Traffic Ground Truth Estimation Using Multisensor Consensus Filter

    MacCarley, C. Arthur / Sandberg, Erik L. | Transportation Research Record | 2012


    HD-Map Based Ground Truth to Test Automated Vehicles

    Sirkar, Deshna / Alzu'bi, Hamzeh / Martens, Joshua et al. | SAE | 2022


    HD-Map Based Ground Truth to Test Automated Vehicles

    Martens, Joshua / Zhong, Ren / Yedida, Siddhartha et al. | British Library Conference Proceedings | 2022