A Novel Agent-Based Approach for Dynamic Emotion Modeling in Social Networks.

Area of Science:

• Social Network Analysis
• Computational Social Science
• Affective Computing

Background:

• Understanding group emotion spread is crucial for identifying social risks in tense environments.
• Traditional propagation models lack individual node behavior and interaction dynamics, introducing randomness.
• Existing models fail to capture the complexities of real-world emotional networks.

Purpose of the Study:

• To develop a novel approach for reconstructing emotion contagion at an individual level.
• To integrate text-based emotion recognition with propagation models for enhanced accuracy.
• To create a deterministic model that reflects individualized node variability in emotion spread.

Main Methods:

• Developed a dynamic agent-based emotion model (AEmo) incorporating multihop agents driven by text emotion analysis.
• Agents record and respond to neighbors' emotional states, enhancing traditional propagation nodes.
• Categorized nodes by emotional states to create corresponding agent types for dynamic modeling.

Main Results:

• The AEmo model reconstructs emotion contagion at an individual level, moving beyond preset probabilities.
• Emotion spread is modeled as a deterministic process with individualized infection rates reflecting node variability.
• Effective prediction of group negative emotion spread and individual emotion evolution was demonstrated on real-world and scale-free networks.

Conclusions:

• The proposed agent-based emotion model (AEmo) offers a more realistic representation of emotion contagion dynamics.
• Integrating text-based emotion recognition provides valuable insights into individual emotion evolution within networks.
• This approach enhances the prediction of social risks by accurately modeling negative emotion spread.