Loading...
Predicting the Polarity of Electronic Word-of-Mouth Communication Created on Tweets Messages Posted by Health Policy Makers Related to Covid-19
Alemi, Mohammad Amin | 2023
84
Viewed
- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 56603 (44)
- University: Sharif University of Technology
- Department: Management and Economics
- Advisor(s): Aslani, Shirin
- Abstract:
- Health influencers leveraged social networks to connect with people and society during the Covid-19 crisis. Platforms like Twitter served as suitable channels for disseminating published messages through word-of-mouth communication. In times of crises like Covid-19, individuals and organizations involved in managing the situation harnessed this capability, contingent upon the quantity and quality of word-of-mouth exchanges. During public crises, public polarity and sentiment toward the issue and the maintenance of public morale hold paramount significance. Consequently, crafting messages that garner positive word-of-mouth communication became a focal point for health influencers. In this research, we endeavored to present a model for predicting the polarity of electronic word-of-mouth communication generated in response to Covid-19-related messages. Our primary data source consisted of tweets from health influencers addressing Covid-19, with the ensuing tweet quotes representing word-of-mouth interactions. We incorporated three categories of tweet characteristics into our model: structural attributes of the tweet itself, features of the tweet's engagement, and elements of the tweet's content. Ultimately, employing machine learning algorithms, our model achieved a 74% accuracy in predicting the average polarity of electronic word-of-mouth communication stemming from a Covid-19-related message. Additionally, we utilized the SHAP model to elucidate the influence of our variables, revealing that metrics such as the number of comments and likes on a tweet, as well as the positive and negative emotional tone of a tweet, exerted the most substantial impact on the average polarity of electronic word-of-mouth communication associated with Covid-19 messages
- Keywords:
- COVID-19 ; Electronic Word to Mouth ; Twitter Social Network ; Public Health Policy Making ; Machine Learning ; Text Polarity ; Text Emotion ; Twitter Posts
- محتواي کتاب
- view