Date of Graduation

12-2020

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Computer Science (PhD)

Degree Level

Graduate

Department

Computer Science & Computer Engineering

Advisor/Mentor

Xiaoqing "Frank" Liu

Committee Member

Brajendra Nath Panda

Second Committee Member

Douglas Adams

Third Committee Member

Xintao Wu

Fourth Committee Member

Susan E. Gauch

Keywords

Cyber Argumentation, Polarization, Quantitative Modeling, Social Connection Recommendation, Stance Prediction

Abstract

Cyber argumentation platforms implement theoretical argumentation structures that promote higher quality argumentation and allow for informative analysis of the discussions. Dr. Liu’s research group has designed and implemented a unique platform called the Intelligent Cyber Argumentation System (ICAS). ICAS structures its discussions into a weighted cyber argumentation graph, which describes the relationships between the different users, their posts in a discussion, the discussion topic, and the various subtopics in a discussion. This platform is unique as it encodes online discussions into weighted cyber argumentation graphs based on the user’s stances toward one another’s arguments and ideas. The resulting weighted cyber argumentation graphs can then be used by various analytical models to measure aspects of the discussion. In prior work, many aspects of cyber argumentation have been modeled and analyzed using these stance relationships.

However, many challenging problems remain in cyber argumentation. In this dissertation, I address three of these problems: 1) modeling and measure argumentation polarization in cyber argumentation discussions, 2) encouraging diverse social networks and preventing echo chambers by injecting ideological diversity into social connection recommendations, and 3) developing a predictive model to predict the stance polarity and intensity relationships between posts in online discussions, allowing discussions from outside of the ICAS platform to be encoded as weighted cyber argumentation graphs and be analyzed by the cyber argumentation models. In this dissertation, I present models to measure polarization in online argumentation discussions, prevent polarizing echo-chambers and diversifying users’ social networks ideologically, and allow online discussions from outside of the ICAS environment to be analyzed using the previous models from this dissertation and the prior work from various researchers on the ICAS system.

This work serves to progress the field of cyber argumentation by introducing a new analytical model for measuring argumentation polarization and developing a novel method of encouraging ideological diversity into social connection recommendations. The argumentation polarization model is the first of its kind to look specifically at the polarization among the users contained within a single discussion in cyber argumentation. Likewise, the diversity enhanced social connection recommendation re-ranking method is also the first of its kind to introduce ideological diversity into social connections. The former model will allow stakeholders and moderators to monitor and respond to argumentation polarization detected in online discussions in cyber argumentation. The latter method will help prevent network-level social polarization by encouraging social connections among users who differ in terms of ideological beliefs. This work also serves as an initial step to expanding cyber argumentation research into the broader online deliberation field. The stance polarity and intensity prediction model presented in this dissertation is the first step in allowing discussions from various online platforms to be encoded into weighted cyber argumentation graphs by predicting the stance weights between users’ posts. These resulting predicted weighted cyber augmentation graphs could then be used to apply cyber argumentation models and methods to these online discussions from popular online discussion platforms, such as Twitter and Reddit, opening many new possibilities for cyber argumentation research in the future.

Download

Share

COinS