• (Wo)man versus machine: An investigation into the quality of automated sentiment analysis, compared to its manual counterpart.

    Author(s):
    Amy Flora Lightfoot (see profile)
    Date:
    2018
    Group(s):
    CityLIS
    Subject(s):
    Library science, Information science
    Item Type:
    Thesis
    Institution:
    City, University of London
    Tag(s):
    library and information, sentiment analysis, Library and information science
    Permanent URL:
    http://dx.doi.org/10.17613/M6XG9F99H
    Abstract:
    "Corporate sentiment analysis is a field in which there is limited public research available. Effective research in this subsection of the sentiment analysis field would help provide a more robust understanding of the field and how users interact online. This study evaluated the efficacy of sentiment analysis using Twitter data containing corporate hashtags. Following the data extraction both a manual classification and an automated one, conducted using the coding language R and the Naive Bayes classifier, were conducted to determine the discrepancy in efficacy between the two. The manual analysis proved more effective and accurate in determining sentiment as well as correcting for spam and irrelevant data. The study, as one of the first of its kind, provides a new avenue of research in sentiment analysis and could encourage researchers to engage with the topic."
    Metadata:
    xml
    Status:
    Published
    Last Updated:
    5 years ago
    License:
    All Rights Reserved
    Share this:

    Downloads

    Item Name: docx woman-versus-machine_-an-investigation-into-the-quality-of-automated-sentiment-analysis-compared-to-its-manual-counterpart.-.docx
      Download
    Activity: Downloads: 165

    Back to Deposits