Modeling the spread of health information using social network analysis : understanding public perception on Dengvaxia

Date of Award

2019

Document Type

Thesis

Degree Name

Master of Science in Computer Science, Straight

Department

Information Systems & Computer Science

First Advisor

Estuar, Ma. Regina Justina E., Ph.D

Abstract

Social media platform like Twitter paved the way for easy information dissemination over the Internet. However, use of social media platform carries high probability of misinformation. In 2018, many parents decided to completely stop getting their children vaccinated due to the Dengvaxia scandal, which resulted to several measles outbreaks. Social media platform contributed to the fast information dissemination regarding the danger of Dengvaxia which created a negative perception towards vaccines in general. The study identified how information regarding the adverse effects of Dengvaxia spread on Twitter. Doc2vec was compared to n-gram neural network classification in order to identify Public Perception on Health Tweets (PPHT). The diffusion characteristics and its corresponding centrality measures was used to model the spread of PPHT and Non-PPHT. The result shows that bigram neural network has the highest performance measure with 85.57% accuracy, 85% precision, 86% recall and 85% F1 score. Moreover, the most influential PPHT comes from Youtube video shares, news agencies and its associates. While influential mediators are users that mostly post tweets to support a particular administration (i.e., Duterte admin). PPHT spreads deeper and has more replies than Non-PPHT, but has a lower structural virality and number of favorites.

Comments

The C7.A27 2019

Link to Full Text

Share

COinS