EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

TRACE: Adaptive Detection and Morphological Classification of Recurring Engagement in YouTube Comment Streams

Dimitrios Zikos, Rodina Bizri-Baryak

Complexity · 2026-01 · 2026 (1)

DOI: 10.1155/cplx/6170840 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Complexity, online erschienen am 17.11.2026. Verfasst von Dimitrios Zikos und 1 weiteren Autor:innen. Die Arbeit nennt 40 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

YouTube videos experience recurring cycles of public engagement where older videos attract renewed attention in response to external events. Traditional event detection frameworks are not designed to capture these reactivation patterns, relying on static thresholds and limited morphological analysis. This study develops the temporal recognition of audience comment engagement (TRACE), a methodology for identifying and classifying phases of public engagement in YouTube comments through retrospective longitudinal analysis, enabling comprehensive examination of aging online content lifecycles. TRACE transforms discrete interaction timestamps into a time‐series dataset, establishes a baseline per video of the expected engagement decay, and extracts periods where observed interaction exceeds that baseline. TRACE incorporates a quiet‐day filtering approach and adaptive residual analysis to account for the natural decay and noise. Detected engagement episodes are morphologically categorized, including spikes, slow burns, and recurring waves, and automatically cross‐referenced against user‐provided lists of dated external events. To demonstrate the algorithm’s versatility across diverse themes, we analyzed three YouTube videos sourced from the health, marketing, and social science domains. These videos were purposively selected for their high historical traction, multiyear persistence, and thematic susceptibility to reactivation via external real‐world events. Across the three domains, TRACE detected hundreds of distinct engagement episodes. Morphological analysis revealed that cyclical engagement patterns (e.g., M‐shaped and triangular) predominated in health controversies and marketing, while politically driven social science content exhibited sharp, isolated spikes. We compared detected engagement episodes against researcher‐composed timelines for each of the three videos to examine associations between the extracted engagement episodes and the researcher‐provided real‐world events (external events) and to demonstrate the cross‐referencing value. The proposed TRACE framework provides a reproducible tool for analyzing the lifecycle and morphology of digital public engagement. It helps challenge assumptions about digital relevance being exclusive to new content by revealing reactivation in archival content. The inclusion of external event matching is not just a detection‐rate step but can also facilitate distinguishing between externally aligned events and unmapped engagement bursts, which can be the foundation for new hypotheses.

Zitieren

APA

Zikos, D., & Bizri-Baryak, R. (2026). TRACE: Adaptive Detection and Morphological Classification of Recurring Engagement in YouTube Comment Streams. Complexity, 2026(1). https://doi.org/10.1155/cplx/6170840

Vancouver

Zikos D, Bizri-Baryak R. TRACE: Adaptive Detection and Morphological Classification of Recurring Engagement in YouTube Comment Streams. Complexity. 2026;2026(1). doi:10.1155/cplx/6170840

BibTeX
@article{Zikos2026,
  author = {Zikos, Dimitrios and Bizri-Baryak, Rodina},
  title = {TRACE: Adaptive Detection and Morphological Classification of Recurring Engagement in YouTube Comment Streams},
  journal = {Complexity},
  year = {2026},
  volume = {2026},
  number = {1},
  doi = {10.1155/cplx/6170840}
}
Kostenlos starten →