Graph-Based Rule Engine for Real-Time Fraud Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing data analysis techniques face challenges in efficiently identifying and tracking relationships among large numbers of entities, particularly in communication networks, where billions of Call Detail Records need to be processed for fraud detection and other applications, requiring scalable and real-time solutions.

Innovation Solution

A real-time graph-based rule engine analyzes connectivity between entities in a graph database, using pre-defined rules to detect changes in relationships and generate notifications, with the database implemented in primary storage (RAM) for ultra-fast response times, enabling the identification of new suspect entities and changes in relationships.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If traditional data analysis techniques are used to process billions of Call Detail Records, then comprehensive relationship analysis can be achieved, but processing time and system response speed increase significantly

Engineering Contradiction:

Improvenumber of entities and relationshipsVSAvoidprocessing time

Core Design Contradiction:

Quantity of substanceVSLoss of time

Solution Approach 1:

The system pre-loads the graph database structure and pre-defines analysis rules before actual processing begins. When new Call Detail Records arrive, the system can immediately query the pre-structured graph without performing comprehensive scans, enabling real-time relationship analysis despite the large volume of data.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent replaces traditional mechanical database scanning and sequential processing with a graph-based query system. By representing entities and relationships as a graph structure with indexed connections, the system can traverse relationships through predefined paths rather than scanning entire datasets, dramatically reducing processing time for large-scale analysis.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Speed

If real-time relationship detection is implemented, then responsive action can be taken, but system complexity and computational requirements increase

Engineering Contradiction:

Improveresponse speedVSAvoidsystem complexity

Core Design Contradiction:

SpeedVSDevice complexity

Solution Approach 1:

The system segments the relationship detection process into distinct components: the graph database stores structured relationship data, the rule engine handles pattern matching, and the notification system manages alerts. This segmentation allows each component to be optimized independently and simplifies the overall architecture by dividing complex analysis tasks into manageable modules.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces a rule engine as an intermediary between the graph database and the notification system. The rule engine receives graph queries, applies predefined relationship patterns, and triggers notifications only when rules are violated. This intermediary layer abstracts the complexity of rule evaluation from the core database operations, enabling real-time response without proportionally increasing system complexity.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Productivity

If graph database stored in RAM is used, then query speed increases, but memory requirements and infrastructure costs increase

Engineering Contradiction:

Improvequery processing speedVSAvoidmemory capacity required

Core Design Contradiction:

ProductivityVSQuantity of substance

Solution Approach 1:

The system loads only the necessary portions of the graph database into RAM based on the specific query requirements. Rather than keeping the entire graph in memory, the system can selectively load relevant entity-relationship sections into RAM for fast querying, while storing the complete graph on disk. This local quality approach maintains high query speeds for active analysis areas while reducing overall memory infrastructure requirements.

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS9154640B2Methods and systems for mass link analysis using rule engines

Publication Date: 2015.10.06 COGNYTE TECH ISRAEL LTD

AI summary

A substantially real-time graph-based rule engine that analyzes connectivities, both direct and indirect relationships, between entities stored in a database as the database is updated (e.g., with CDR or financial transaction data). The rule engine uses pre-defined rules to detect events (i.e., the database updates) that influence the relationship between entities in the database. When the database is updated with events (e.g., CDRs), the real-time rule engine compares the update to any relevant rules. If the real-time based rule engine finds a match between a rule and an update to the database, then the rule engine generates a notification, such as an alert. The alerts may be used to provide notification of, e.g, fraudulent activities.