University of Illinois Urbana-Champaign

A structured matrix factorization method for computational modeling of hierarchical polarization in social interactions

Sun, Dachun

This item is only available for download by members of the University of Illinois community. Students, faculty, and staff at the U of I may log in with your NetID and password to view the item. If you are trying to access an Illinois-restricted dissertation or thesis, you can request a copy through your library's Inter-Library Loan office or purchase a copy directly from ProQuest.

Permalink

https://hdl.handle.net/2142/120391

Description

Title
A structured matrix factorization method for computational modeling of hierarchical polarization in social interactions
Author(s)
Sun, Dachun
Issue Date
2023-04-20
Director of Research (if dissertation) or Advisor (if thesis)
Abdelzaher, Tarek
Department of Study
Computer Science
Discipline
Computer Science
Degree Granting Institution
University of Illinois at Urbana-Champaign
Degree Name
M.S.
Degree Level
Thesis
Keyword(s)
  • Polarization
  • Belief estimation
  • Hierarchical
  • Matrix factorization
  • Unsupervised
Abstract
Many works on social interaction polarization detection focus heavily on flat classification of stances and beliefs. We extend them in this work in two important aspects: (i) detects both points of agreement and disagreement between groups, and (ii) divides them hierarchically to represent nested patterns of agreement and disagreement given a structural guide. For example, two opposing parties might disagree on core issues. Moreover, a disagreement might occur on further details within a party, despite agreement on the fundamentals. We call such scenarios hierarchically polarization. An unsupervised Non-negative Matrix Factorization (NMF) algorithm is described for the computational modeling of hierarchical polarization in social interactions. The algorithm is enhanced with a language model and a proof of orthogonality of factorized components. We evaluate it on both synthetic and real-world datasets, demonstrating the ability to decompose overlapping beliefs hierarchically. In the case where polarization is flat, we compare it to the prior art and show that it outperforms state-of-the-art approaches for polarization detection and stance separation. An ablation study further illustrates the value of individual components, including new enhancements.
Graduation Semester
2023-05
Type of Resource
Thesis
Copyright and License Information
Copyright 2023 Dachun Sun

Owning Collections

Graduate Dissertations and Theses at Illinois PRIMARY
Graduate Theses and Dissertations at Illinois