Fraud Detection Algorithm Dynamic Parameter Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems fail to effectively detect and handle false positives in fraudulent online transactions, leading to errors that compromise the quality of service provided by financial organizations, as they often misidentify legitimate transactions as fraudulent.
Innovation Solution
A system and method that analyze user actions and malware actions data using a hardware processor with a predetermined algorithm, adjusting operating parameters to minimize false positives by calculating and updating frame values and cost functions based on user interactions during electronic transactions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional systems use fixed algorithms to detect fraudulent transactions, then detection consistency is maintained, but false positives increase and service quality deteriorates
Solution Approach 1:
The patent implements dynamic adjustment of algorithm operating parameters based on analyzed user behavior patterns. The system transitions from static fixed algorithms to dynamic adaptive algorithms that modify their parameters in real-time based on detected user interactions, thereby improving detection precision while maintaining reliability through continuous adaptation to legitimate user behaviors.
Solution Approach 2:
The system changes operating parameters of the detection algorithm based on analyzed user behavior data. By adjusting parameters such as detection thresholds and time window sizes according to observed user patterns, the system reduces false positives while maintaining effective fraud detection, directly addressing the precision-reliability contradiction.
2Measurement precision
If the system analyzes detailed user actions data, then detection accuracy improves, but processing time and computational complexity increase
Solution Approach 1:
The patent segments user behavior analysis into specific actionable parameters rather than processing all raw data uniformly. By dividing the analysis into targeted segments such as click patterns, navigation sequences, and interaction timing, the system achieves high detection accuracy while reducing overall processing time through focused analysis of critical behavior segments.
Solution Approach 2:
The system performs partial analysis by focusing on the most discriminative user behavior features rather than exhaustive analysis of all possible data points. This selective approach to analyzing only the most relevant user actions maintains high detection accuracy while significantly reducing computational overhead and processing time.
3Device complexity
If the system uses static algorithm parameters, then system simplicity is maintained, but false positive rates increase
Solution Approach 1:
The patent introduces dynamic parameter adjustment mechanisms that adapt algorithm operating parameters based on analyzed user behavior patterns. This dynamic approach increases service quality by reducing false positives through adaptive detection, while the underlying algorithm structure remains relatively simple, requiring only basic parameter adjustment capabilities rather than complete structural redesign.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Disclosed is a system and method for detecting fraudulent transactions. An example method includes receiving data relating to an electronic transaction, including at least one of user actions data and malware actions data; analyzing, the data to determine whether the electronic transaction is a possible fraudulent transaction based on a predetermined algorithm stored in an electronic memory; determining whether the possible fraudulent transaction is a legitimate electronic transaction; and adjusting the operating parameters of the predetermined algorithm if the hardware processor determines that the possible fraudulent transaction is a legitimate electronic transaction.