Lakehead University Library Logo
    • Login
    View Item 
    •   Knowledge Commons
    • Electronic Theses and Dissertations
    • Electronic Theses and Dissertations from 2009
    • View Item
    •   Knowledge Commons
    • Electronic Theses and Dissertations
    • Electronic Theses and Dissertations from 2009
    • View Item
    JavaScript is disabled for your browser. Some features of this site may not work without it.
    quick search

    Browse

    All of Knowledge CommonsCommunities & CollectionsBy Issue DateAuthorTitleSubjectDisciplineAdvisorCommittee MemberThis CollectionBy Issue DateAuthorTitleSubjectDisciplineAdvisorCommittee Member

    My Account

    Login

    Statistics

    View Usage Statistics

    Local and Global Influence on Twitter

    Thumbnail

    View/Open

    ZongS2016m-1a.pdf (1.509Mb)

    Date

    2016

    Author

    Zong, Shan

    Degree

    Master of Science

    Discipline

    Engineering : Electrical & Computer

    Subject

    Influence analysis
    Social network service
    2D classification

    Metadata

    Show full item record

    Abstract

    The analysis of influence in social network is drawing more and more attention. It can be applied in different areas such as political campaigns and marketing. In this work, the analysis of influence in Twitter, based on users‟ profile statistics in a real-time scale, was studied and discussed. Two methods of identifying influential users by given keyword in real-time are introduced. To understand the relationship between users‟ influence features and social states in real life, two influence measures were presented: Local Influence which the user has on his/her immediate set of contacts and global Influence which the user has on the entire social network. This study describes in details these two metrics and shows their implementation for a real social network. Our case study, using Twitter, showed that the proposed model can create clusters of users in 2D space corresponding to their social standing, and can further be used to classify previously-unseen users into the correct classes with an f-measure of 0.82 which is significantly higher than benchmark algorithms. F-measure is often used for measuring the accuracy of the test for classification.

    URI

    http://knowledgecommons.lakeheadu.ca/handle/2453/813

    Collections

    • Electronic Theses and Dissertations from 2009

    Lakehead University Library
    Contact Us | Send Feedback

     


    Lakehead University Library
    Contact Us | Send Feedback