A Simultaneous Extraction of Context and Community from pervasive signals using nested Dirichlet process

Thuong Nguyen (Corresponding Author), Vu Nguyen, Flora Salim, L Duc Le Viet Duc, Dinh Phung

    Research output: Contribution to journalArticleAcademicpeer-review

    Abstract

    Understanding user contexts and group structures plays a central role in pervasive computing. These contexts and community structures are complex to mine from data collected in the wild due to the unprecedented growth of data, noise, uncertainties and complexities. Typical existing approaches would first extract the latent patterns to explain human dynamics or behaviors and then use them as a way to consistently formulate numerical representations for community detection, often via a clustering method. While being able to capture high-order and complex representations, these two steps are performed separately. More importantly, they face a fundamental difficulty in determining the correct number of latent patterns and communities. This paper presents an approach that seamlessly addresses these challenges to simultaneously discover latent patterns and communities in a unified Bayesian nonparametric framework. Our Simultaneous Extraction of Context and Community (SECC) model roots in the nested Dirichlet process theory which allows a nested structure to be built to summarize data at multiple levels. We demonstrate our framework on five datasets where the advantages of the proposed approach are validated.
    Original languageEnglish
    Pages (from-to)396-417
    Number of pages22
    JournalPervasive and Mobile Computing
    Volume38
    DOIs
    Publication statusPublished - Jul 2017

    Fingerprint

    Ubiquitous computing
    Uncertainty

    Cite this

    @article{db933b23ee30448db6a8186a071b6cc5,
    title = "A Simultaneous Extraction of Context and Community from pervasive signals using nested Dirichlet process",
    abstract = "Understanding user contexts and group structures plays a central role in pervasive computing. These contexts and community structures are complex to mine from data collected in the wild due to the unprecedented growth of data, noise, uncertainties and complexities. Typical existing approaches would first extract the latent patterns to explain human dynamics or behaviors and then use them as a way to consistently formulate numerical representations for community detection, often via a clustering method. While being able to capture high-order and complex representations, these two steps are performed separately. More importantly, they face a fundamental difficulty in determining the correct number of latent patterns and communities. This paper presents an approach that seamlessly addresses these challenges to simultaneously discover latent patterns and communities in a unified Bayesian nonparametric framework. Our Simultaneous Extraction of Context and Community (SECC) model roots in the nested Dirichlet process theory which allows a nested structure to be built to summarize data at multiple levels. We demonstrate our framework on five datasets where the advantages of the proposed approach are validated.",
    author = "Thuong Nguyen and Vu Nguyen and Flora Salim and {Le Viet Duc}, {L Duc} and Dinh Phung",
    year = "2017",
    month = "7",
    doi = "10.1016/j.pmcj.2016.08.019",
    language = "English",
    volume = "38",
    pages = "396--417",
    journal = "Pervasive and Mobile Computing",
    issn = "1574-1192",
    publisher = "Elsevier",

    }

    A Simultaneous Extraction of Context and Community from pervasive signals using nested Dirichlet process. / Nguyen, Thuong (Corresponding Author); Nguyen, Vu; Salim, Flora; Le Viet Duc, L Duc; Phung, Dinh.

    In: Pervasive and Mobile Computing, Vol. 38, 07.2017, p. 396-417.

    Research output: Contribution to journalArticleAcademicpeer-review

    TY - JOUR

    T1 - A Simultaneous Extraction of Context and Community from pervasive signals using nested Dirichlet process

    AU - Nguyen, Thuong

    AU - Nguyen, Vu

    AU - Salim, Flora

    AU - Le Viet Duc, L Duc

    AU - Phung, Dinh

    PY - 2017/7

    Y1 - 2017/7

    N2 - Understanding user contexts and group structures plays a central role in pervasive computing. These contexts and community structures are complex to mine from data collected in the wild due to the unprecedented growth of data, noise, uncertainties and complexities. Typical existing approaches would first extract the latent patterns to explain human dynamics or behaviors and then use them as a way to consistently formulate numerical representations for community detection, often via a clustering method. While being able to capture high-order and complex representations, these two steps are performed separately. More importantly, they face a fundamental difficulty in determining the correct number of latent patterns and communities. This paper presents an approach that seamlessly addresses these challenges to simultaneously discover latent patterns and communities in a unified Bayesian nonparametric framework. Our Simultaneous Extraction of Context and Community (SECC) model roots in the nested Dirichlet process theory which allows a nested structure to be built to summarize data at multiple levels. We demonstrate our framework on five datasets where the advantages of the proposed approach are validated.

    AB - Understanding user contexts and group structures plays a central role in pervasive computing. These contexts and community structures are complex to mine from data collected in the wild due to the unprecedented growth of data, noise, uncertainties and complexities. Typical existing approaches would first extract the latent patterns to explain human dynamics or behaviors and then use them as a way to consistently formulate numerical representations for community detection, often via a clustering method. While being able to capture high-order and complex representations, these two steps are performed separately. More importantly, they face a fundamental difficulty in determining the correct number of latent patterns and communities. This paper presents an approach that seamlessly addresses these challenges to simultaneously discover latent patterns and communities in a unified Bayesian nonparametric framework. Our Simultaneous Extraction of Context and Community (SECC) model roots in the nested Dirichlet process theory which allows a nested structure to be built to summarize data at multiple levels. We demonstrate our framework on five datasets where the advantages of the proposed approach are validated.

    UR - https://authors.elsevier.com/a/1VBrU5bwSmkCWH

    U2 - 10.1016/j.pmcj.2016.08.019

    DO - 10.1016/j.pmcj.2016.08.019

    M3 - Article

    VL - 38

    SP - 396

    EP - 417

    JO - Pervasive and Mobile Computing

    JF - Pervasive and Mobile Computing

    SN - 1574-1192

    ER -