Lakehead University Library Logo
    • Login
    View Item 
    •   Knowledge Commons Home
    • Electronic Theses and Dissertations
    • Electronic Theses and Dissertations from 2009
    • View Item
    •   Knowledge Commons Home
    • 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 DateAuthorsTitlesSubjectsDisciplineAdvisorCommittee MemberThis CollectionBy Issue DateAuthorsTitlesSubjectsDisciplineAdvisorCommittee Member

    My Account

    Login

    Fairness, engagement, and discourse analysis in AI-driven social media and healthcare

    Thumbnail
    View/Open
    SinghalA2023m-1a.pdf (4.957Mb)
    Date
    2023
    Author
    Singhal, Aditya
    Metadata
    Show full item record
    Abstract
    This thesis addresses the critical concerns of fairness, accountability, transparency, and ethics (FATE) within the context of artificial intelligence (AI) systems applied to social media and healthcare domains. First, a comprehensive survey examines existing research on FATE in AI, specifically focusing on the subdomains of social media and healthcare. The survey evaluates current solutions, highlights their benefits, limitations, and potential challenges, and charts out future research directions. Key findings emphasize the significance of statistical and intersectional fairness in ensuring equitable healthcare access on social media platforms and highlight the pivotal role of transparency in AI systems to foster accountability. Building upon the survey, this thesis delves into an analysis of social media usage by healthcare organizations, with a specific emphasis on engagement and sentiment forecasting during the COVID-19 pandemic. Data collection from Twitter handles of pharmaceutical companies, public health agencies, and the World Health Organization enables extensive analysis. Natural language processing (NLP)-based topic modeling techniques are applied to identify health-related topics, while sentiment forecasting models are employed to gauge public sentiment. The results uncover the impact of COVID-19-related topics on public engagement, highlighting the varying levels of engagement across diverse healthcare organizations. Notably, the World Health Organization exhibits dynamic engagement patterns over time, necessitating adaptable strategies. The thesis further presents latest sentiment forecasting models, such as autoregressive integrated moving average (ARIMA) and seasonal autoregressive integrated moving average with exogenous factors (SARIMAX), which enable organizations to optimize their content strategies for maximum user engagement. Furthermore, discourse analysis is conducted to unravel the factors that shape the content of tweets by healthcare organizations on Twitter. [...]
    URI
    https://knowledgecommons.lakeheadu.ca/handle/2453/5204
    Collections
    • Electronic Theses and Dissertations from 2009 [1632]

    Lakehead University Library
    Contact Us | Send Feedback

     

     


    Lakehead University Library
    Contact Us | Send Feedback