Integrating Node Features and Structure: Distance Measures in Social Network Comparison

This work approaches network clustering in the context of similar networks of YouTube chan-nels for food-related content by using network analysis. We use optimal transport algorithms to calculate the distance across networks while taking similarity in structure and content into account. Subsequently, channels with comparable viewership trends and content overlap are grouped using spectral clustering. The generated clusters highlight different channel groups that demonstrate thematic content alignment and shared audience targeting tactics. This makes it feasible for marketers, platform managers, and content producers to make insightful deductions about potential joint ventures, cross-promotions, and market expansion. The study demonstrated how useful network-based and optimal transport distance techniques can be for comprehending and structuring digital content ecosystems.