What is it about?
This study examines how internet memes rise, fade, or remain popular over time. Using Google Trends data from 2,126 memes, machine learning identified four common popularity patterns: rapid decline, long-term growth, stable interest after a peak, and decline with occasional resurgences. A support vector classification model recognized these patterns with about 98% accuracy. The study also found that meme types, such as catchphrases, characters, templates, pop-culture references, and viral videos, are associated with different popularity lifecycles. Overall, the research provides a scalable way to understand how online content spreads and remains culturally relevant.
Featured Image
Photo by David Valentine on Unsplash
Why is it important?
This study expands earlier research from about 100 memes to more than 2,000 and introduces an automated machine learning method for studying their popularity. Instead of focusing only on whether content becomes viral, it examines what happens after the peak. This is timely because memes now influence communication, entertainment, marketing, and public discussion. The framework may help researchers and content professionals better understand short-term attention and long-term cultural influence.
Perspectives
This publication is meaningful to me because it connects mathematical modeling, machine learning, and digital culture. Memes may appear to be simple entertainment, but their spread reflects important patterns of human attention and cultural transmission. I also value the interdisciplinary and student-centered nature of the project. I see this study as a foundation for future work on early prediction, cross-platform spread, and the effects of content and context on long-term popularity.
Pengcheng Xiao
Kennesaw State University
Read the Original
This page is a summary of: Temporal pattern classification of internet meme propagation: A hybrid machine learning approach, Electronic Research Archive, January 2025, Tsinghua University Press,
DOI: 10.3934/era.2025238.
You can read the full text:
Contributors
The following have contributed to this page
