Link Analysis Noise Filtering via Entity Growth Rate Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for analyzing communication traffic in cellular networks struggle to accurately identify relationships between entities, often generating false relationships due to noisy entities that communicate with unrelated parties, leading to erroneous results in applications like fraud detection and social network analysis.
Innovation Solution
A method that uses call detail records (CDRs) to identify and disqualify noisy entities by assessing their growth rate and communication patterns, constructing a data structure that represents genuine relationships while disregarding false interrelations, and maintaining a 'black list' of noisy entities to improve the reliability of relationship identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If communication traffic is analyzed to identify relationships between entities, then relationship detection capability is improved, but false relationships are generated due to noisy entities
Solution Approach 1:
The patent extracts and identifies noisy entities from the communication network by analyzing growth rates of entity sets. Once identified, these noisy entities are removed from further relationship analysis, preventing them from generating false relationships. This extraction principle directly addresses the contradiction by eliminating the source of false relationships while preserving genuine relationship detection capabilities.
Solution Approach 2:
The patent performs preliminary identification and disqualification of noisy entities before conducting relationship analysis. By evaluating growth rates and identifying noisy entities in advance, the system prevents false relationships from being generated during the relationship detection process. This preliminary action ensures that only reliable entity relationships are analyzed, improving both accuracy and reliability.
2Quantity of substance
If all entities are used in relationship analysis, then analysis coverage is improved, but noise from unrelated communication increases
Solution Approach 1:
The patent converts the harmful effect of noisy entities into a beneficial identification process. By analyzing communication patterns and growth rates, the system identifies noisy entities and uses this information to improve relationship analysis. The noise itself becomes a signal for identification, allowing the system to distinguish between genuine and false relationships while maintaining comprehensive analysis coverage.
Solution Approach 2:
The patent introduces an intermediary filtering mechanism that evaluates growth rates of entity sets. This intermediary process identifies noisy entities and separates them from genuine relationships, allowing the system to maintain high analysis coverage while eliminating noise. The growth rate evaluation acts as a mediator between comprehensive entity inclusion and noise reduction.
3Measurement precision
If growth rate evaluation is performed to identify noisy entities, then relationship accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent changes the parameter used for entity evaluation from simple communication frequency to growth rate of entity sets. This parameter change enables more accurate identification of noisy entities by capturing dynamic changes in communication patterns. The growth rate parameter provides a more sophisticated measure that improves relationship identification accuracy while maintaining computational feasibility through efficient evaluation methods.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
Systems and methods for identifying and characterizing relationships based on communication traffic. The methods may include accepting indications of communication conducted among entities over a communication network, and using a link processor, analyzing the indications so as to identify that two or more of the entities are interrelated by detecting one or more intermediate entities with which the two or more entities communicate. A criterion may be evaluated with respect to the indications associated with a given entity, responsively to meeting the criterion, disqualifying the given entity from serving as an intermediate entity in analyzing the indications.