A pagerank-inspired heuristic scheme for influence maximization in social networks

Bo Zhang, Yufeng Wang, Qun Jin, Jianhua Ma

    Research output: Contribution to journalArticle

    3 Citations (Scopus)

    Abstract

    This article focused on seeking a new heuristic algorithm for the influence maximization problem in complex social networks, in which a small subset of individuals are intentionally selected as seeds to trigger a large cascade of further adoptions of a new behavior under certain influence cascade models. In literature, degree and other centrality-based heuristics are commonly used to estimate the influential power of individuals in social networks. The major issues with degree-based heuristics are twofold. First, those results are only derived for the uniform IC model, in which propagation probabilities on all social links are set as same, which is rarely the case in reality; Second, intuitively, an individual's influence power depends not only on the number of direct friends, but also relates to kinds of those friends, that is, the neighbors' influence should also be taken into account when measuring one's influential power. Based on the general weighted cascade model (WC), this article proposes Pagerank-inspired heuristic scheme, PRDiscount, which explicitly discounts the influence power of those individuals who have social relationships with chosen seeds, to alleviate the "overlapping effect" occurred in behavior diffusion. Then, the authors use both the artificially constructed social network graphs (with the features of power-law degree distribution and small-world characteristics) and the real-data traces of social networks to verify the performance of their proposal. Simulations illustrate that PRDiscount can advantage over the existing degree-based discount algorithm, DegreeDiscount, and achieve the comparable performance as greedy algorithm.

    Original languageEnglish
    Pages (from-to)48-62
    Number of pages15
    JournalInternational Journal of Web Services Research
    Volume12
    Issue number4
    DOIs
    Publication statusPublished - 2015 Oct 1

    Fingerprint

    Seed
    Heuristic algorithms

    Keywords

    • Heuristic
    • Influence maximization
    • Overlapping effect
    • Pagerank
    • Word-of-mouth

    ASJC Scopus subject areas

    • Computer Networks and Communications
    • Software
    • Information Systems

    Cite this

    A pagerank-inspired heuristic scheme for influence maximization in social networks. / Zhang, Bo; Wang, Yufeng; Jin, Qun; Ma, Jianhua.

    In: International Journal of Web Services Research, Vol. 12, No. 4, 01.10.2015, p. 48-62.

    Research output: Contribution to journalArticle

    @article{a5fcca3e9327497c84a7408570d5bbd7,
    title = "A pagerank-inspired heuristic scheme for influence maximization in social networks",
    abstract = "This article focused on seeking a new heuristic algorithm for the influence maximization problem in complex social networks, in which a small subset of individuals are intentionally selected as seeds to trigger a large cascade of further adoptions of a new behavior under certain influence cascade models. In literature, degree and other centrality-based heuristics are commonly used to estimate the influential power of individuals in social networks. The major issues with degree-based heuristics are twofold. First, those results are only derived for the uniform IC model, in which propagation probabilities on all social links are set as same, which is rarely the case in reality; Second, intuitively, an individual's influence power depends not only on the number of direct friends, but also relates to kinds of those friends, that is, the neighbors' influence should also be taken into account when measuring one's influential power. Based on the general weighted cascade model (WC), this article proposes Pagerank-inspired heuristic scheme, PRDiscount, which explicitly discounts the influence power of those individuals who have social relationships with chosen seeds, to alleviate the {"}overlapping effect{"} occurred in behavior diffusion. Then, the authors use both the artificially constructed social network graphs (with the features of power-law degree distribution and small-world characteristics) and the real-data traces of social networks to verify the performance of their proposal. Simulations illustrate that PRDiscount can advantage over the existing degree-based discount algorithm, DegreeDiscount, and achieve the comparable performance as greedy algorithm.",
    keywords = "Heuristic, Influence maximization, Overlapping effect, Pagerank, Word-of-mouth",
    author = "Bo Zhang and Yufeng Wang and Qun Jin and Jianhua Ma",
    year = "2015",
    month = "10",
    day = "1",
    doi = "10.4018/IJWSR.2015100104",
    language = "English",
    volume = "12",
    pages = "48--62",
    journal = "International Journal of Web Services Research",
    issn = "1545-7362",
    publisher = "IGI Publishing",
    number = "4",

    }

    TY - JOUR

    T1 - A pagerank-inspired heuristic scheme for influence maximization in social networks

    AU - Zhang, Bo

    AU - Wang, Yufeng

    AU - Jin, Qun

    AU - Ma, Jianhua

    PY - 2015/10/1

    Y1 - 2015/10/1

    N2 - This article focused on seeking a new heuristic algorithm for the influence maximization problem in complex social networks, in which a small subset of individuals are intentionally selected as seeds to trigger a large cascade of further adoptions of a new behavior under certain influence cascade models. In literature, degree and other centrality-based heuristics are commonly used to estimate the influential power of individuals in social networks. The major issues with degree-based heuristics are twofold. First, those results are only derived for the uniform IC model, in which propagation probabilities on all social links are set as same, which is rarely the case in reality; Second, intuitively, an individual's influence power depends not only on the number of direct friends, but also relates to kinds of those friends, that is, the neighbors' influence should also be taken into account when measuring one's influential power. Based on the general weighted cascade model (WC), this article proposes Pagerank-inspired heuristic scheme, PRDiscount, which explicitly discounts the influence power of those individuals who have social relationships with chosen seeds, to alleviate the "overlapping effect" occurred in behavior diffusion. Then, the authors use both the artificially constructed social network graphs (with the features of power-law degree distribution and small-world characteristics) and the real-data traces of social networks to verify the performance of their proposal. Simulations illustrate that PRDiscount can advantage over the existing degree-based discount algorithm, DegreeDiscount, and achieve the comparable performance as greedy algorithm.

    AB - This article focused on seeking a new heuristic algorithm for the influence maximization problem in complex social networks, in which a small subset of individuals are intentionally selected as seeds to trigger a large cascade of further adoptions of a new behavior under certain influence cascade models. In literature, degree and other centrality-based heuristics are commonly used to estimate the influential power of individuals in social networks. The major issues with degree-based heuristics are twofold. First, those results are only derived for the uniform IC model, in which propagation probabilities on all social links are set as same, which is rarely the case in reality; Second, intuitively, an individual's influence power depends not only on the number of direct friends, but also relates to kinds of those friends, that is, the neighbors' influence should also be taken into account when measuring one's influential power. Based on the general weighted cascade model (WC), this article proposes Pagerank-inspired heuristic scheme, PRDiscount, which explicitly discounts the influence power of those individuals who have social relationships with chosen seeds, to alleviate the "overlapping effect" occurred in behavior diffusion. Then, the authors use both the artificially constructed social network graphs (with the features of power-law degree distribution and small-world characteristics) and the real-data traces of social networks to verify the performance of their proposal. Simulations illustrate that PRDiscount can advantage over the existing degree-based discount algorithm, DegreeDiscount, and achieve the comparable performance as greedy algorithm.

    KW - Heuristic

    KW - Influence maximization

    KW - Overlapping effect

    KW - Pagerank

    KW - Word-of-mouth

    UR - http://www.scopus.com/inward/record.url?scp=84954556542&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=84954556542&partnerID=8YFLogxK

    U2 - 10.4018/IJWSR.2015100104

    DO - 10.4018/IJWSR.2015100104

    M3 - Article

    AN - SCOPUS:84954556542

    VL - 12

    SP - 48

    EP - 62

    JO - International Journal of Web Services Research

    JF - International Journal of Web Services Research

    SN - 1545-7362

    IS - 4

    ER -