Integrating Individual Factors to Construct Recognition Models of Consumer Fraud Victimization

Area of Science:

• Social Sciences
• Criminology
• Behavioral Economics

Background:

• Consumer financial fraud poses significant economic, physical, mental, social, and legal risks to victims.
• Proactive identification of vulnerable individuals can help mitigate the pervasive threat of financial scams.

Purpose of the Study:

• To develop and evaluate machine learning models for predicting consumer fraud exposure and victimhood.
• To identify key individual factors influencing susceptibility to financial fraud within the Chinese population.

Main Methods:

• Utilized a nationwide survey of 36,202 participants within a two-stage conceptual framework.
• Employed machine learning techniques to construct Fraud Exposure Recognition (FER) and Fraud Victimhood Recognition (FVR) models.
• Interpreted model components to identify significant predictive factors for fraud exposure.

Main Results:

• The FER model demonstrated strong predictive performance with an f1 score of 0.727.
• Key predictors for fraud exposure included migration status, financial status, urbanicity, and age in the Chinese context.
• The FVR model exhibited lower predictive power (f1 = 0.565), suggesting the need for additional psychological variables.

Conclusions:

• Individual factors like demographics and socioeconomic status are significant predictors of financial fraud exposure.
• Future research on consumer fraud victimhood should incorporate psychological dimensions for improved predictive accuracy.
• This study offers valuable insights for understanding individual differences in vulnerability to consumer financial fraud.