Table 1 The explanation of risk perception.

General risk perception

Different people’s subjective judgment or appraisal of risk

Questionnaires, interviews, psychometrics, and other qualitative methods

(e.g., collecting people’s views on certain risk events, having them fill out questionnaires based on subjective perceptions, or scoring are largely psychological interpretations. These methods tend to be highly subjective and prone to inaccuracies due to variations in individual perceptions.)

Data science methods, such as machine learning, natural language processing, and statistical techniques like linear regression, can also predict risk.

(e.g., the semantic vector approach, compared to standard methods in risk perception studies, involves regressing risk source ratings on nine correlation dimensions.)

Tourism risk perception

Different tourist’s recognition and evaluation of risk during traveling

People’s a priori knowledge from hearsay or preconceptions about a destination (e.g., asking travelers for opinions on certain risk events is often inaccurate. Variations in expression and the listener’s ability to grasp information mean risk size is judged solely on others’ descriptions, ignoring population heterogeneity.)

Text mining approaches based on website data (e.g., constructing a risk perception measure from travel notes of past visitors, synthesizing public perceptions and individual feelings about specific risks, and providing an objective and efficient reference for future travelers can offer precise insights.)