"Agent Evaluation via Social Network Analysis in Educational E-CARGO Framework"

"The data set includes 10 years (from 2014 to 2023) of course enrollment data, the data about links between teachers and courses. Artificial intelligence technology is being increasingly applied in-depth to education. Smart education exemplifies agent collaboration, involving negotiation, evaluation, allocation, and execution among agents in distinct roles. The social network of agents and environmental elements significantly influences such collaboration. This paper first formalizes the process as a Role-Based Collaboration (RBC) workflow and describes a real-world scenario. Subsequently, we define an E-CARGO model for education that incorporates a social network structure. As the core contribution, the application of Social Network Analysis (SNA) to agent evaluation in educational RBC is investigated for the first time. A suite of algorithms is defined and evaluated using real-world Teacher-Course-Student (TCS) data. Experimental results reveal that path-based SNA algorithms substantially outperform existing baselines, particularly when data accumulates over time, demonstrating great potential of SNA for agent evaluation."