Fraudulent Message Detection Using Reference Point Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional spam filtering software fails to effectively detect fraudulent 'phishing' electronic messages that appear to come from reputable sources, leading to potential personal information theft from unsuspecting users.
Innovation Solution
A system and process that uses reference points in messages to classify them, identifying divergent reference points and applying statistical analysis to detect fraudulent messages, including the use of a fraud detection engine that examines messages for indicators such as raw IP addresses, non-standard encoding, and forms requesting personal information, and alerts users to potential phishing attempts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional spam filtering software is used to detect fraudulent messages, then junk email can be filtered, but fraudulent phishing messages that appear legitimate cannot be effectively detected
Solution Approach 1:
The patent segments message analysis into multiple components: extracting reference points (URLs, email addresses, phone numbers), classifying each reference point's legitimacy, identifying divergent reference points (mismatches between claimed and actual sources), and performing statistical analysis. This segmentation allows the system to detect subtle fraudulent indicators that conventional filters miss while maintaining efficiency.
Solution Approach 2:
The patent adds a new dimension of analysis by examining the relationship between reference points and their claimed sources. Instead of just filtering based on keywords or sender reputation, the system analyzes whether reference points (URLs, email addresses) actually match the purported sender identity, creating a multi-dimensional verification approach that detects phishing messages disguised as legitimate communications.
2Productivity
If spam filtering software places messages in spam folder, then junk email is separated, but legitimate fraudulent messages may be misclassified and recipients may still respond
Solution Approach 1:
The patent performs preliminary analysis by extracting and classifying reference points before final message classification. By identifying divergent reference points early in the process and performing statistical analysis on message characteristics, the system can flag fraudulent messages for further review or direct quarantine, preventing them from reaching users even when they appear legitimate.
Solution Approach 2:
The system implements feedback mechanisms where detected fraudulent messages and their characteristics are used to refine detection rules and improve future classification accuracy. The statistical analysis component learns from patterns in fraudulent messages, continuously improving detection precision while maintaining filtering efficiency.
3Reliability
If users manually review messages to detect fraud, then detection accuracy improves, but time consumption and user burden increase significantly
Solution Approach 1:
The patent implements an automated self-service detection system that performs reference point extraction, classification, and statistical analysis without user intervention. The system autonomously identifies divergent reference points, evaluates message legitimacy, and takes appropriate actions (quarantine, delete, or deliver), eliminating time-consuming manual review while maintaining high detection reliability through multi-layered automated analysis.
Data Source
AI summary
A technique for classifying a message is disclosed. In some embodiments, the technique comprises extracting a plurality of reference points, classifying the plurality of reference points, and detecting that the message is a phish message based on the classified reference points. In some embodiments, the technique comprises identifying a plurality of fraud indicators in the message, applying a statistical analysis on the plurality of fraud indicators; and determining whether the message is a fraudulent message based on the analysis.