Fraud Detection Strategy Ranking to Eliminate Overlapping Alerts

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing fraud detection systems face inefficiencies due to overlapping and duplicate fraud alerts, leading to resource-intensive processing and ineffective flagging of fraudulent transactions, with limited operational power and difficulty in evaluating each rule's net contribution.

Innovation Solution

A computing device, the fraud optimizer, uses a customized forward selection algorithm to evaluate and rank fraud detection strategies by calculating a fraud value for each transaction, removing overlaps, and generating a ranked list to optimize resource use and improve fraud detection efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If multiple fraud detection strategies are applied to all transactions, then comprehensive fraud detection coverage is achieved, but processing time and computational resources increase significantly

Engineering Contradiction:

Improvefraud detection coverageVSAvoidprocessing time

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system pre-calculates fraud values for each transaction under each strategy and pre-ranks strategies based on their historical performance and overlap characteristics. This preliminary preparation allows the system to quickly apply pre-ranked strategies to new transactions without re-evaluating all strategies from scratch, significantly reducing processing time while maintaining comprehensive detection coverage

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent segments the fraud detection process into distinct phases: (1) applying strategies in ranked order, (2) identifying and removing overlapping transactions after each strategy application, (3) re-ranking remaining strategies based on updated fraud values, and (4) repeating the process. This segmentation allows the system to process transactions in manageable stages rather than evaluating all strategies simultaneously, reducing computational burden while maintaining detection effectiveness

Inventive Principle:

Principle #1Segmentation

2Reliability

If all fraud detection strategies process all transactions, then no fraudulent transactions are missed, but computational resources are wasted on duplicate and low-value alerts

Engineering Contradiction:

Improvefraud detection accuracyVSAvoidcomputational resources

Core Design Contradiction:

ReliabilityVSLoss of energy

Solution Approach 1:

The system applies fraud detection strategies in a partial manner by processing transactions through strategies in ranked order and stopping when a transaction is flagged by a high-priority strategy. Not all strategies need to evaluate all transactions - once a transaction is identified as fraudulent by a top-ranked strategy, lower-ranked strategies can skip that transaction. This partial action approach maintains detection accuracy while significantly reducing computational resource consumption

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The system discards transactions that have already been flagged as fraudulent by higher-priority strategies, removing them from the pool of transactions to be evaluated by subsequent strategies. This prevents duplicate processing of the same fraudulent transaction across multiple strategies. The discarded transactions are effectively recovered from further processing, eliminating wasted computational resources on already-identified fraud cases

Inventive Principle:

Principle #34Discarding and recovering

3Productivity

If fraud detection strategies are ranked by highest fraud value, then resource allocation is optimized, but the complexity of evaluating and ranking strategies increases

Engineering Contradiction:

Improveresource allocation efficiencyVSAvoidevaluation and ranking complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The system automatically evaluates and ranks fraud detection strategies based on their performance metrics and overlap characteristics without requiring manual intervention. The ranking mechanism self-adjusts as transactions are processed, with strategies being re-ranked based on updated fraud values and transaction removals. This self-service approach handles the complexity internally while presenting a simplified interface to users, maintaining resource allocation efficiency without burdening operators with complex evaluation tasks

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS12481997B2System and method for optimization of fraud detection model

Publication Date: 2025.11.25 THE TORONTO DOMINION BANK

AI summary

There is provided a computing system for optimizing a plurality of fraud detection strategies used to generate a corresponding set of potentially fraudulent transactions. The system determines an overall fraud value such as an average fraud value for each transaction based on pre-defined factors and identifies a particular strategy having a highest average fraud value for its fraudulent transactions as a highest priority on a ranked list of strategies. The system is configured to remove each transaction from the remaining other strategies if the same as the fraudulent transactions in the identified strategy and calculate an average fraud value for the remaining other strategies. The system then ranks the next highest priority fraud detection strategy having the highest average fraud value while removing its corresponding transactions flagged from other remaining strategies and repeat the ranking until all the strategies have been ranked and apply the ranked list to subsequent transactions.