Fraud Detection System Analyzing User Action Series

The patent merges multiple separate operation assessments into a unified fraud detection framework. The fraud determination score is calculated by integrating information from a series of operations (login, screen transitions, user information access) rather than evaluating each operation in isolation. This combining approach allows the system to capture the contextual relationships between operations, thereby improving fraud detection accuracy while maintaining manageable system complexity through structured integration.

The patent implements preliminary action by pre-defining the detection framework and score calculation methodology before actual fraud detection occurs. The system establishes the relationship between operations and fraud indicators in advance, allowing it to efficiently process user behavior sequences without requiring complex real-time analysis. This preparation enables the system to maintain fast processing speeds while comprehensively analyzing behavioral context.

The patent applies segmentation by dividing the fraud detection process into distinct operational stages (login detection, screen transition monitoring, user information access tracking) while integrating them through a unified scoring mechanism. Each operation type is analyzed with appropriate depth, and the segmented results are combined to form an overall fraud assessment. This segmentation approach maintains detection reliability for individual operations while enabling comprehensive fraud detection across the entire user journey.