Transaction Screening System Reducing False Positives
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Financial institutions face challenges in detecting suspicious transactions while minimizing false positives, given the large volume of transactions that precludes 100% human review.
Innovation Solution
A computer system that aggregates transactions for individual clients across their accounts, applying screening criteria to detect unusual patterns, such as multiple large withdrawals, and flags accounts for human review while excluding internal transactions and regular activities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If computer systems automatically screen all transactions to detect suspicious activity, then the detection capability is improved, but the number of false positives increases
Solution Approach 1:
The patent segments the large volume of transactions by organizing them into client-level groups and applying screening criteria at multiple levels (transaction level, account level, and client level). This segmentation allows the system to detect suspicious patterns more effectively while reducing false positives by contextualizing transactions within client behavior profiles.
Solution Approach 2:
The system dynamically adjusts screening parameters and criteria based on client-specific patterns and historical behavior. By changing the parameters of what constitutes 'suspicious' activity based on individual client profiles, the system improves detection accuracy while reducing false alarms that would occur with rigid, one-size-fits-all screening rules.
2Measurement precision
If all transactions are reviewed by human analysts, then detection accuracy is improved, but the productivity and efficiency decrease
Solution Approach 1:
The system performs preliminary automated screening of all transactions before human review, pre-processing the data to identify and flag only those transactions that meet suspicious criteria. This preliminary action filters out the majority of normal transactions, allowing human analysts to focus their expertise on a smaller, more relevant subset of transactions, thereby maintaining high detection accuracy while improving overall productivity.
3Reliability
If screening criteria are set to detect all questionable transactions, then the detection coverage is improved, but the volume of transactions requiring human review increases excessively
Solution Approach 1:
The patent applies different screening criteria and thresholds at different levels of the transaction hierarchy. Rather than using uniform criteria for all transactions, the system tailors the screening approach to local characteristics such as client type, account history, and transaction patterns. This local quality approach ensures comprehensive detection coverage while adjusting the sensitivity to minimize the volume of transactions that require human review.
Data Source
AI summary
A computer system detects questionable financial transactions. The system includes a data source component configured to access a database of transaction data, and a data storage component configured to store transaction data that is to be reviewed. An account compilation component of the system is configured to compile data sets from the transaction data. Each data set includes pertinent external transactions in the transaction data that are related to a particular individual client during a screening time interval. The account compilation component excludes internal transactions from the data sets. A screening component is configured to compare each data set with at least one screening criterion. A reporting component is configured to report data sets that satisfy the at least one screening criterion.