Abstract:
This study applies a hybrid configurational approach to investigate SIM box fraud patterns in telecommunications data. By integrating behavioral indicators derived from call detail records with Necessary Condition Analysis (NCA) and Fuzzy-set Qualitative Comparative Analysis (fsQCA), the research identifies the conditions and specific combinations of call characteristics that lead to fraudulent outcomes. After excluding the fields that restate the outcome, the NCA results show that international call type is a fully necessary condition for SIM box fraud (effect size = 1.000) and that short call duration imposes a near-complete constraint (effect size = 0.978), while no other recorded attribute bounds the outcome. The analysis demonstrates that fraud behavior does not stem from isolated factors but rather emerges from complex configurations of conditions such as call duration, call type, and origin–destination patterns.
The findings underscore the interpretive value of configurational analysis in telecom fraud detection compared with purely predictive, black-box models. While predictive approaches often obscure the underlying mechanisms of fraud, the fsQCA framework reveals how multiple behavioral conditions interact to generate fraudulent activity. This methodological contribution enhances transparency, supports actionable insights for fraud management, and provides a foundation for bridging explanatory and predictive approaches in fraud detection research.
