Risk perception and behavioral change during epidemics: Comparing models of individual and collective learning.

Area of Science:

• Socio-environmental systems
• Risk perception and diffusion
• Agent-based modeling

Background:

• Societies face diverse risks, necessitating understanding of risk awareness spread and coping strategy diffusion.
• Individual risk perception and communication about protective measures are driven by learning and social interaction.
• Agent-based modeling (ABM) is used to study diffusion, incorporating context-dependent learning and social interactions, often combined with machine learning.

Purpose of the Study:

• To investigate the role of intelligent learning in risk appraisal and protective decision-making.
• To explore the differences between individual and collective learning within diffusion models, specifically in socio-environmental systems.
• To analyze the impact of intelligent learning on the spectrum from individual to collective learning using an enhanced agent-based model.

Main Methods:

• Development of an agent-based model (ABM) incorporating machine learning for intelligent learning.
• Simulation experiments to compare different learning and decision-making strategies (individual, majority vote, leader-based).
• Analysis of the influence of social interactions on both individual and group learning processes.

Main Results:

• Individual intelligent judgment and group majority voting on risks and coping strategies were less effective than leader-based groups or individuals deciding alone.
• Social interactions were found to be crucial for both individual and group learning processes.
• The representation of social learning in ABMs should consider prevailing cultural and social norms.

Conclusions:

• Intelligent learning, particularly in leader-based structures, can enhance risk appraisal and protective decision-making compared to simple majority rule.
• Social interactions are fundamental drivers for effective learning in socio-environmental risk diffusion.
• Future agent-based models should integrate cultural and social norms to accurately represent social learning dynamics.