Fraud Detection System Using Multi-Source Risk Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current fraud detection systems face challenges in identifying procurement-related fraud and risk due to the elusive nature of fraudulent activities, high false positive rates, and the difficulty in tracing responsible parties, with existing methods often requiring labeled data and being inefficient in processing unstructured data.
Innovation Solution
A system and method that analyze standard transactional data from multiple public and private sources, utilizing a learning component with formal guarantees to compute vendor and requestor risk scores, and an active invoice score, which integrates text analytics, social network analysis, and sequential probabilistic learning to identify fraudulent entities with a low false positive rate.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional fraud detection methods are used, then detection capability is limited, but false positive rate increases and detection accuracy decreases
Solution Approach 1:
The patent combines multiple data sources including public data, private data, transactional data, and unstructured data into a unified analysis framework. This integration allows the system to cross-validate information from multiple channels, improving detection accuracy while reducing false positives through corroborating evidence from diverse sources.
Solution Approach 2:
The system transitions from traditional single-dimension fraud detection to multi-dimensional analysis by incorporating social network analysis, entity relationship mapping, and contextual information from multiple data sources. This dimensional expansion enables more precise fraud identification by examining patterns across different analytical planes.
2Measurement precision
If comprehensive data analysis is performed to improve detection accuracy, then detection capability improves, but system complexity increases
Solution Approach 1:
The patent segments the fraud detection system into distinct functional modules: data collection module, data processing module, scoring module, and investigation module. Each module handles specific tasks independently, managing complexity through functional decomposition while maintaining comprehensive analysis capabilities across multiple data sources and analytical techniques.
Solution Approach 2:
The system introduces intermediary components including data normalization layers, scoring algorithms, and risk assessment models that mediate between raw multi-source data and final fraud determination. These intermediaries simplify the complexity by providing structured transformation and interpretation layers that handle the integration of diverse data formats and analytical results.
3Measurement precision
If multiple data sources are integrated to provide comprehensive analysis, then detection capability improves, but processing time increases
Solution Approach 1:
The system performs preliminary actions by pre-collecting and pre-processing data from multiple sources before fraud detection is triggered. Data normalization, entity resolution, and initial risk assessment are conducted in advance, allowing the system to quickly evaluate new transactions against pre-computed profiles and patterns, reducing real-time processing requirements.
Solution Approach 2:
The patent implements continuous data collection and analysis processes that operate backgroundingly across multiple data sources. Rather than batch-processing all data at once, the system continuously monitors and updates entity profiles, maintaining ready-to-use analytical results that can be quickly applied to new fraud detection scenarios without requiring complete re-analysis.
Data Source
AI summary
An apparatus, method and computer program product for identifying fraud in transaction data. The method includes: receiving invoice data comprising a vendor, a requestor and events, receiving public data and private data sources, computing a vendor risk score using the public and private data sources matching the vendor of the invoice data, computing a requestor risk score using the public data sources and the private data sources matching the requestor of the invoice data, computing an active invoice score using the vendor risk score and the requestor risk score and when the active invoice score is greater than a predetermined amount, blocking the invoice data. In one embodiment, computing a vendor risk score comprises obtaining a weight and a confidence for the event, calculating an event vendor risk score using the weight times the confidence and combining the event vendor risk scores for all of the events.