A two-phase model of collective memory decay with a dynamical switching point

Area of Science:

• Social Sciences
• Computational Social Science
• Information Science

Background:

• Collective memory, the shared recall of significant events, has been studied since the 1920s.
• Digitization and online data enable quantitative analysis of collective memory dynamics.
• Previous models often used single decay processes and limited event categories.

Purpose of the Study:

• To propose and validate a novel two-phase mathematical model for collective memory decay.
• To investigate collective memory dynamics across diverse event categories using empirical data.
• To identify patterns and switching points in memory decay processes.

Main Methods:

• Developed a two-phase mathematical model combining exponential (fast) and power-law (slow) decay.
• Applied the model to Wikipedia page view data for five event categories: earthquakes, deaths, aviation accidents, mass murders, and terrorist attacks.
• Compared the proposed model's performance against existing collective memory decay models.

Main Results:

• The proposed two-phase model demonstrated superior performance compared to existing models across most event categories.
• Estimated model parameters were consistent across all analyzed event categories.
• A dynamical switching point from exponential to power-law decay was consistently detected around 10-11 days post-event peak.

Conclusions:

• The two-phase model effectively captures the complex dynamics of collective memory decay.
• Memory decay dynamics exhibit consistent patterns across different types of significant societal events.
• The identified switching point provides a quantifiable measure of collective memory persistence.