Fraud Detection System Using Preliminary User Action Screening

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing fraud detection systems take a long time to complete processing after detecting a detection target due to the need for fraud detection, which prolongs the time before predetermined processing can be executed.

Innovation Solution

A fraud detection system that determines whether a predetermined action has been performed by a user before detection, executes fraud detection based on identification information stored in the user terminal, and then performs predetermined processing upon detection, thereby shortening the overall time required.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If fraud detection is executed after detection target is detected, then fraud detection accuracy is ensured, but processing time is excessively long

Engineering Contradiction:
Improvefraud detection accuracyVSAvoidprocessing time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system executes fraud detection in advance by monitoring predetermined actions (such as code generation, copying, or distribution actions) before the detection target is actually used. This preliminary fraud detection based on user behavior patterns allows the system to pre-determine fraud risk, so that when the detection target is subsequently detected, the processing can be completed quickly using the pre-obtained fraud determination result, thereby resolving the contradiction between ensuring detection accuracy and reducing processing time

Inventive Principle:
Principle #10Preliminary action

2Loss of time

If fraud detection is executed before detection target is detected, then processing time is shortened, but system complexity increases

Engineering Contradiction:
Improveprocessing timeVSAvoidsystem complexity
Core Design Contradiction:
Loss of timeVSDevice complexity

Solution Approach 1:

The system performs preliminary fraud detection by monitoring specific predetermined actions (code generation, copying, distribution) that are naturally occurring in the workflow. This approach avoids the need for continuous comprehensive monitoring, thereby limiting the increase in system complexity to only the extent necessary for detecting these specific actions, while still achieving the benefit of shortened processing time

Inventive Principle:
Principle #10Preliminary action

3Reliability

If continuous monitoring is performed to detect predetermined actions, then fraud detection accuracy is improved, but energy consumption increases

Engineering Contradiction:
Improvefraud detection accuracyVSAvoidenergy consumption
Core Design Contradiction:
ReliabilityVSUse of energy by moving object

Solution Approach 1:

Instead of continuous monitoring, the system performs periodic detection by monitoring only when specific predetermined actions occur (code generation, copying, or distribution actions). This event-driven approach ensures high fraud detection accuracy by focusing on critical moments while significantly reducing energy consumption by avoiding constant monitoring, thus resolving the contradiction between reliability and energy usage

Inventive Principle:
Principle #19Periodic action

Data Source

PatentUS12456122B2Fraud detection system, fraud detection device, fraud detection method, and program
Publication Date: 2025.10.28 RAKUTEN GROUP INC
  • US12456122B2 patent drawing
  • US12456122B2 patent drawing
  • US12456122B2 patent drawing

AI summary

A fraud detection system for executing predetermined processing when a detection target is detected by using a detection device, the fraud detection system comprising at least one processor which determines, before the detection target is detected, whether a predetermined action has been performed by a user having a user terminal; executes, when it is determined that the predetermined action has been performed, fraud detection on the user based on identification information stored in the user terminal; and executes, when the detection target is detected, the predetermined processing based on an execution result of the fraud detection.