The 16th Annual International Conference on Advances in Social Networks Analysis and Mining (ASONAM) leads data mining interdisciplinary research. As the name implies, the conference focuses on Social Network Analysis and Mining (SNAM), which has gained importance in various fields such as academia, business, politics, and homeland security and has its roots in social and business communities. This research—driven largely by the analysis of online platforms, email records, phone logs, and instant messaging systems—commonly uses graph theory and machine learning. It mirrors the broader trend of society evolving into networks where individuals rely more heavily on their connections.
ASONAM 2024 aimed to offer a cross-disciplinary platform for researchers and practitioners in different areas of SNAM to collaborate and share ideas. The event focused on key issues, emphasizing emerging trends and the demands of the industry. The conference invited both theoretical and experimental research on social network analysis and mining, with a focus on real-world applications.
From September 2 to 5, 2024, the 16th ASONAM conference was held at the University of Calabria in Calabria, Italy. This year COSMOS had 7 papers that were accepted and published in the proceedings of the conference by Springer! Several cosmographers visited the conference to present their research. The following is a list of papers from cosmographers that were published in the conference:
- Are Narratives Contagious? Modeling Narrative Diffusion Using Epidemiological Theories
- Beyond the Click: How YouTube Thumbnails Shape User Interaction and Algorithmic Recommendations
- Detecting and Measuring Anomalous Behaviors on YouTube
- Mitigating the Spread of COVID-19 Misinformation Using Agent-Based Modeling and Delays in Information Diffusion
- Multi-agent Analytics-Driven Content Discovery: A Narrative Contagion Approach
- Navigating the Anomalies: A Comprehensive Analysis of YouTube Channel Behavior
- Utilizing Fractional Order Epidemiological Model to Understand High and Moderate Toxicity Spread on Social Media Platforms
In this first article of a two-part series, we summarize three of these papers:
- Navigating the Anomalies: A Comprehensive Analysis of YouTube Channel Behavior
- Detecting and Measuring Anomalous Behaviors on YouTube
- Beyond the Click: How YouTube Thumbnails Shape User Interaction and Algorithmic Recommendations
The former two studies focus on anomalous interaction trends on YouTube that can indicate artificially boosted or misleading content, and all three focus on interaction from YouTube users with trends in videos.
“Navigating the Anomalies: A Comprehensive Analysis of YouTube Channel Behavior” used engagement metrics and commenter behavior to identify anomalous YouTube channels through Cosine Similarity and Principal Component Analysis (PCA). Cosine Similarity calculated the resemblance between active and suspended channels, while PCA identified key features explaining dataset variance. Findings revealed that anomalies vary across metrics, requiring a multi-dimensional approach.
“Detecting and Measuring Anomalous Behaviors on YouTube” extended this by adding a qualitative analysis of anomalous channels. It highlighted trends in contentious content and distinct network properties, with key indicators including clustering coefficient, comment volume, and unique commenter count.
“Beyond the Click” analyzed how YouTube thumbnails influence user interaction and algorithmic recommendations. Attributes like brightness, colorfulness, and quality showed minimal impact on engagement for sensitive topics. Instead, engagement was driven by content relevance, with YouTube’s algorithm favoring visually engaging thumbnails, which can marginalize minority viewpoints and promote content homogenization.