Estimating Personal Network Size with Non-random Mixing via Latent Kernels

Swupnil Sahai, Timothy Jones, Sarah Cowan, Tian Zheng

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Abstract

    A major problem in the study of social networks is estimating the number of people an individual knows. However, there is no general method to account for barrier effects, a major source of bias in common estimation procedures. The literature describes approaches that model barrier effects, or non-random mixing, but they suffer from unstable estimates and fail to give results that agree with specialists’ knowledge. In this paper we introduce a model that builds off existing methods, imposes more structure, requires significantly fewer parameters, and yet allows for greater interpretability. We apply our model on responses gathered from a survey we designed and show that our conclusions better match what sociologists find in practice. We expect that this approach will provide more accurate estimates of personal network sizes and hence remove a significant hurtle in sociological research.

    Original languageEnglish (US)
    Title of host publicationComplex Networks and Their Applications VII - Volume 1 Proceedings The 7th International Conference on Complex Networks and their Applications COMPLEX NETWORKS 2018
    EditorsRenaud Lambiotte, Luis M. Rocha, Pietro Lió, Hocine Cherifi, Luca Maria Aiello, Chantal Cherifi
    PublisherSpringer-Verlag
    Pages694-705
    Number of pages12
    ISBN (Print)9783030054106
    DOIs
    StatePublished - Jan 1 2019
    Event7th International Conference on Complex Networks and their Applications, COMPLEX NETWORKS 2018 - Cambridge, United Kingdom
    Duration: Dec 11 2018Dec 13 2018

    Publication series

    NameStudies in Computational Intelligence
    Volume812
    ISSN (Print)1860-949X

    Other

    Other7th International Conference on Complex Networks and their Applications, COMPLEX NETWORKS 2018
    CountryUnited Kingdom
    CityCambridge
    Period12/11/1812/13/18

    Keywords

    • Barrier effects
    • Kernel-based models
    • Non-random mixing
    • Personal network size estimation

    ASJC Scopus subject areas

    • Artificial Intelligence

    Cite this

    Sahai, S., Jones, T., Cowan, S., & Zheng, T. (2019). Estimating Personal Network Size with Non-random Mixing via Latent Kernels. In R. Lambiotte, L. M. Rocha, P. Lió, H. Cherifi, L. M. Aiello, & C. Cherifi (Eds.), Complex Networks and Their Applications VII - Volume 1 Proceedings The 7th International Conference on Complex Networks and their Applications COMPLEX NETWORKS 2018 (pp. 694-705). (Studies in Computational Intelligence; Vol. 812). Springer-Verlag. https://doi.org/10.1007/978-3-030-05411-3_55

    Estimating Personal Network Size with Non-random Mixing via Latent Kernels. / Sahai, Swupnil; Jones, Timothy; Cowan, Sarah; Zheng, Tian.

    Complex Networks and Their Applications VII - Volume 1 Proceedings The 7th International Conference on Complex Networks and their Applications COMPLEX NETWORKS 2018. ed. / Renaud Lambiotte; Luis M. Rocha; Pietro Lió; Hocine Cherifi; Luca Maria Aiello; Chantal Cherifi. Springer-Verlag, 2019. p. 694-705 (Studies in Computational Intelligence; Vol. 812).

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Sahai, S, Jones, T, Cowan, S & Zheng, T 2019, Estimating Personal Network Size with Non-random Mixing via Latent Kernels. in R Lambiotte, LM Rocha, P Lió, H Cherifi, LM Aiello & C Cherifi (eds), Complex Networks and Their Applications VII - Volume 1 Proceedings The 7th International Conference on Complex Networks and their Applications COMPLEX NETWORKS 2018. Studies in Computational Intelligence, vol. 812, Springer-Verlag, pp. 694-705, 7th International Conference on Complex Networks and their Applications, COMPLEX NETWORKS 2018, Cambridge, United Kingdom, 12/11/18. https://doi.org/10.1007/978-3-030-05411-3_55
    Sahai S, Jones T, Cowan S, Zheng T. Estimating Personal Network Size with Non-random Mixing via Latent Kernels. In Lambiotte R, Rocha LM, Lió P, Cherifi H, Aiello LM, Cherifi C, editors, Complex Networks and Their Applications VII - Volume 1 Proceedings The 7th International Conference on Complex Networks and their Applications COMPLEX NETWORKS 2018. Springer-Verlag. 2019. p. 694-705. (Studies in Computational Intelligence). https://doi.org/10.1007/978-3-030-05411-3_55
    Sahai, Swupnil ; Jones, Timothy ; Cowan, Sarah ; Zheng, Tian. / Estimating Personal Network Size with Non-random Mixing via Latent Kernels. Complex Networks and Their Applications VII - Volume 1 Proceedings The 7th International Conference on Complex Networks and their Applications COMPLEX NETWORKS 2018. editor / Renaud Lambiotte ; Luis M. Rocha ; Pietro Lió ; Hocine Cherifi ; Luca Maria Aiello ; Chantal Cherifi. Springer-Verlag, 2019. pp. 694-705 (Studies in Computational Intelligence).
    @inproceedings{4ffdb27de34f4fdf9d5cf5b963ad0838,
    title = "Estimating Personal Network Size with Non-random Mixing via Latent Kernels",
    abstract = "A major problem in the study of social networks is estimating the number of people an individual knows. However, there is no general method to account for barrier effects, a major source of bias in common estimation procedures. The literature describes approaches that model barrier effects, or non-random mixing, but they suffer from unstable estimates and fail to give results that agree with specialists’ knowledge. In this paper we introduce a model that builds off existing methods, imposes more structure, requires significantly fewer parameters, and yet allows for greater interpretability. We apply our model on responses gathered from a survey we designed and show that our conclusions better match what sociologists find in practice. We expect that this approach will provide more accurate estimates of personal network sizes and hence remove a significant hurtle in sociological research.",
    keywords = "Barrier effects, Kernel-based models, Non-random mixing, Personal network size estimation",
    author = "Swupnil Sahai and Timothy Jones and Sarah Cowan and Tian Zheng",
    year = "2019",
    month = "1",
    day = "1",
    doi = "10.1007/978-3-030-05411-3_55",
    language = "English (US)",
    isbn = "9783030054106",
    series = "Studies in Computational Intelligence",
    publisher = "Springer-Verlag",
    pages = "694--705",
    editor = "Renaud Lambiotte and Rocha, {Luis M.} and Pietro Li{\'o} and Hocine Cherifi and Aiello, {Luca Maria} and Chantal Cherifi",
    booktitle = "Complex Networks and Their Applications VII - Volume 1 Proceedings The 7th International Conference on Complex Networks and their Applications COMPLEX NETWORKS 2018",

    }

    TY - GEN

    T1 - Estimating Personal Network Size with Non-random Mixing via Latent Kernels

    AU - Sahai, Swupnil

    AU - Jones, Timothy

    AU - Cowan, Sarah

    AU - Zheng, Tian

    PY - 2019/1/1

    Y1 - 2019/1/1

    N2 - A major problem in the study of social networks is estimating the number of people an individual knows. However, there is no general method to account for barrier effects, a major source of bias in common estimation procedures. The literature describes approaches that model barrier effects, or non-random mixing, but they suffer from unstable estimates and fail to give results that agree with specialists’ knowledge. In this paper we introduce a model that builds off existing methods, imposes more structure, requires significantly fewer parameters, and yet allows for greater interpretability. We apply our model on responses gathered from a survey we designed and show that our conclusions better match what sociologists find in practice. We expect that this approach will provide more accurate estimates of personal network sizes and hence remove a significant hurtle in sociological research.

    AB - A major problem in the study of social networks is estimating the number of people an individual knows. However, there is no general method to account for barrier effects, a major source of bias in common estimation procedures. The literature describes approaches that model barrier effects, or non-random mixing, but they suffer from unstable estimates and fail to give results that agree with specialists’ knowledge. In this paper we introduce a model that builds off existing methods, imposes more structure, requires significantly fewer parameters, and yet allows for greater interpretability. We apply our model on responses gathered from a survey we designed and show that our conclusions better match what sociologists find in practice. We expect that this approach will provide more accurate estimates of personal network sizes and hence remove a significant hurtle in sociological research.

    KW - Barrier effects

    KW - Kernel-based models

    KW - Non-random mixing

    KW - Personal network size estimation

    UR - http://www.scopus.com/inward/record.url?scp=85059081011&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=85059081011&partnerID=8YFLogxK

    U2 - 10.1007/978-3-030-05411-3_55

    DO - 10.1007/978-3-030-05411-3_55

    M3 - Conference contribution

    SN - 9783030054106

    T3 - Studies in Computational Intelligence

    SP - 694

    EP - 705

    BT - Complex Networks and Their Applications VII - Volume 1 Proceedings The 7th International Conference on Complex Networks and their Applications COMPLEX NETWORKS 2018

    A2 - Lambiotte, Renaud

    A2 - Rocha, Luis M.

    A2 - Lió, Pietro

    A2 - Cherifi, Hocine

    A2 - Aiello, Luca Maria

    A2 - Cherifi, Chantal

    PB - Springer-Verlag

    ER -