A Review of Entropy-Based Studies on Crowd Behavior and Risk Analysis

Understanding crowd behavior has become imperative as the number of mass gatherings is increasing. Many factors, such as overcrowding, venue deficiencies, rumors, accidents (such as fires, etc.), and inadequate crowd management practices, can lead to serious crowd-related severe that have resulted in injuries and fatalities. Several studies have been conducted to understand crowd behavior and predict risk conditions using different methods. The methods used to detect the crowd anomaly behavior can be classified into two general categories; Implicit Method (relies on expert opinions on several factors and needs human intervention in the model) and Explicit Method (Assessment of crowd risk situations done by the model itself, requires less human intervention in the model). Entropy, one of such explicit methods, has been widely adopted to analyze crowd behavior and predict crowd risk situations in recent years. The purpose of this study is to review the existing literature on entropy-based crowd risk prediction models and identify research gaps so that more in-depth studies on crowd management and risk assessment can be conducted.