Skip to content

Latest commit

 

History

History
52 lines (36 loc) · 2.75 KB

README.md

File metadata and controls

52 lines (36 loc) · 2.75 KB

Structural-Polarization-Twitter

Final Research Project - Complex and Social Networks (CSN)

Authors: Sara Montese, Marius Behret

Table of Contents

  1. Domain
  2. Aims of the Research
  3. Hypotheses to be Tested
  4. Theoretical Framework
  5. Methods
  6. Report
  7. Code

rt_israel edges

Domain

Structural polarization refers to the phenomenon where a network (such as a social or political network) becomes divided into distinct groups or “poles” based on certain opinions. In the context of the paper “Separating Polarization from Noise: Comparison and Normalization of Structural Polarization Measures” [1], structural polarization refers to the quantification of polarization within a network structure.

Aims of the Research

  • Analyze structural polarization of datasets from Social Media
  • Compare the results of standard and normalized polarization measures
  • Study the correlation of the polarization measures under analysis
  • Propose one or more polarization scores

Hypotheses to be Tested

  • Randomized networks appear to be polarized even if they are randomly generated using the eight polarization measures from [1]
  • The choice of the polarization scores has a lower impact than their normalization
  • The process of normalizing polarization scores effectively reduces significant interference caused by the local properties of the network

Theoretical Framework

Network science, social network analysis, polarization measurement, community detection, and clustering.

Methods

Generation of the randomized non-polarized networks; calculation of the several polarization scores like Random Walk Controversy (RWC), Adaptive Random Walk Controversy (ARWC), Betweenness Centrality Controversy (BCC), Boundary Polarization (BP), Dipole polarization (DP), E-I Index (EI), Adaptive E-I Index (AEI) and Modularity (Q) on polarized and non-polarized networks. Comparison between standard and normalized polarization metrics.

Report

For more detailed information, please refer to our research report here.

Code

Find the code implementation in the 'code' directory.

References

  1. A. Salloum, T. H. Y. Chen, and M. Kivelä, “Separating polarization from noise: comparison and normalization of structural polarization measures,” Proceedings of the ACM on human-computer interaction, vol. 6, no. CSCW1, pp. 1–33, 2022

Acknowledgments

This research is conducted at the Facultat d'Informàtica de Barcelona (FIB), Universitat Politècnica de Catalunya (UPC) - BarcelonaTech.