Unmasking Machine Learning With Tensor Decomposition: An Illustrative Example for Media and Communication Researchers

Aug 4, 2026

Online communication datasets keep expanding, while manual content analysis used in social-science media studies struggles with scalability, efficiency, and limited coding scope. Automated machine learning can improve scale but often behaves as a black box, providing little interpretability into which features drive predictions. This study highlights explainable AI for balancing accuracy and transparency, yet adoption in communication research remains limited. It demonstrates tensor decomposition—PARAFAC2—as an interpretable approach for analyzing high-dimensional communication data.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages20

Reading Time50 min

AccessFree to download

Abstract

1. Introduction

[www.ssoar. info](www.ssoar. info) Unmasking Machine Learning With Tensor Decomposition: An Illustrative Example for Media and Communication Researchers Oh , Yu Won; Park , Chong Hyun Veröffentlichungsversion / Published Version Zeitschriftenartikel / journal article Empfohlene Zitierung / Suggested Citation: Oh , Y. W. , & Park , C. H. (2025) . Unmasking Machine Learning With Tensor Decomposition: An Illustrative Example for Media and Communication Researchers. Media and Communication, 13. [https://doi.org/10.17645/mac.9623](https://doi.org/10.17645/mac.9623) Nutzungsbedingungen: Dieser Text wird unter einer CC BY Lizenz (Namensnennung) zur Verfügung gestellt. Nähere Auskünfte zu den CC-Lizenzen finden Sie hier: [https://creativecommons.org/licenses/by/4.0/deed.de](https://creativecommons.org/licenses/by/4.0/deed.de) Terms of use: This document is made available under a CC BY Licence (Attribution). For more Information see: [https://creativecommons.org/licenses/by/4.0](https://creativecommons.org/licenses/by/4.0) Media and Communication 2025 • Volume 13 • Article 9623 [https://doi.org/10.17645/mac.9623](https://doi.org/10.17645/mac.9623) |
ARTICLE | Open Access Journal
| | --- | --- | | Unmasking Machine Learning With Tensor Decomposition: An Illustrative Example for Media and Communication Researchers
Yu Won Oh 1 and Chong Hyun Park 2
1 School of Digital Media, Myongji University, Republic of Korea
2 School of Business, Sungkyunkwan University, Republic of Korea
Correspondence: Chong Hyun Park ([chypark@skku.edu](chypark@skku.edu))
Submitted: 15 November 2024 Accepted: 5 February 2025 Published: 24 April 2025
Issue: This article is part of the issue “AI, Media, and People: The Changing Landscape of User Experiences and Behaviors” edited by Jeong‐Nam Kim (University of Oklahoma) and Jaemin Jung (Korea Advanced Institute of Science and Technology), fully open access at [https://doi.org/10.17645/mac.i475](https://doi.org/10.17645/mac.i475) | | | Abstract
As online communication data continues to grow, manual content analysis, which is frequently employed in media studies within the social sciences, faces challenges in terms of scalability, efficiency, and coding scope Automated machine learning can address these issues, but it often functions as a black box, offering little insight into the features driving its predictions. This lack of interpretability limits its application in advancing social science communication research and fostering practical outcomes. Here, explainable AI offers a solution that balances high prediction accuracy with interpretability. However, its adoption in social science communication studies remains limited. This study illustrates tensor decomposition—specifically, PARAFAC2—for media scholars as an interpretable machine learning method for analyzing high‐dimensional communication data. By transforming complex datasets into simpler components, tensor decomposition reveals the nuanced relationships among linguistic features. Using a labeled spam review dataset as an illustrative example, this study demonstrates how the proposed approach uncovers patterns overlooked by traditional methods and enhances insights into language use. This framework bridges the gap between accuracy and explainability, offering a robust tool for future social science communication research.
Keywords
automated content analysis; explainable AI; machine learning; PARAFAC2; tensor decomposition | | | 1. Introduction
As vast amounts of communication data accumulate online, manual content analysis, which is widely employed in media studies within the social sciences, faces limitations in terms of coding scope, effort, and efficiency | | © 2025 by the author(s), licensed under a Creative Commons Attribution 4 .0 International License (CC BY) . 1 (Kroon et al., 2024) . This explains why automated approaches—despite suspicions of being “incorrect models of l

FAQs

Why is manual content analysis difficult for online media and communication research?

It faces limitations in scalability, efficiency, and the scope of coding for large, continuously growing communication data.

What problem does automated machine learning introduce in this research context?

It often functions as a black box, so the features driving predictions remain unclear and the model lacks interpretability.

How does the study use tensor decomposition to improve interpretability?

It illustrates PARAFAC2 to transform high-dimensional communication data into simpler components, revealing nuanced relationships among linguistic features that traditional methods may miss.

Aug 4, 2026

Upload by River Wang