Graph Data Structure Building for Correlated App Network Views

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing systems struggle to efficiently build and utilize graph data structures from fragmented, inconsistently named, and under-documented raw data to provide coherent and enhanced views of entity networks, particularly in accounting platforms.

Innovation Solution

A method and system for building a graph data structure by determining vertices and edges from raw data, inferring relationships, and updating the graph incrementally using builder modules and a graph interface module, which includes event-driven and batch processing to create a knowledge graph that represents entity networks.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If graph data structures are built from fragmented and inconsistently named raw data, then coherent views of entity networks can be created, but the complexity of data processing and normalization increases significantly

Engineering Contradiction:

Improvecoherence of entity network viewVSAvoiddata processing complexity

Core Design Contradiction:

Loss of informationVSDevice complexity

Solution Approach 1:

The patent segments the graph building process into distinct modules: a graph interface module that handles data ingestion and normalization, builder modules that construct graph structures from normalized data, and separate processing pathways for event-driven and batch operations. This modular segmentation isolates the complexity of data normalization within the interface module while presenting simplified operations to higher-level components.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The graph interface module serves as an intermediary layer between the raw fragmented data and the graph building processes. It performs data normalization, validation, and transformation, converting inconsistently named raw data into a standardized format that downstream builder modules can process efficiently, thereby shielding the complex normalization logic from the rest of the system.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Reliability

If incremental updates are implemented to maintain graph data structures, then the system can provide up-to-date insights, but the overhead of continuous processing increases

Engineering Contradiction:

Improvedata freshnessVSAvoidprocessing overhead

Core Design Contradiction:

ReliabilityVSUse of energy by stationary object

Solution Approach 1:

The system implements dynamic processing modes that can adapt between event-driven incremental updates and batch processing. The graph interface module receives real-time events and can trigger incremental graph updates when necessary, while also supporting scheduled batch processing for less time-sensitive operations, allowing the system to dynamically adjust its processing intensity based on operational needs.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The patent incorporates periodic batch processing alongside event-driven updates. Builder modules can be configured to process accumulated data changes in periodic batches, reducing the frequency of intensive processing operations while maintaining data freshness. This periodic action allows the system to balance between continuous responsiveness and processing overhead by consolidating updates at intervals.

Inventive Principle:

Principle #19Periodic action

3Measurement precision

If correlation analysis is performed on application pairs to generate recommendations, then accurate insights into application relationships are provided, but the computational requirements increase

Engineering Contradiction:

Improvecorrelation accuracyVSAvoidcomputational power

Core Design Contradiction:

Measurement precisionVSPower

Solution Approach 1:

The system implements correlation analysis selectively rather than computing all possible application pair correlations. The builder modules focus computation on identifying and analyzing correlations that are most relevant to the specific domain context, performing partial correlation analysis on subsets of application pairs based on usage patterns and relationships, thereby reducing overall computational requirements while maintaining accuracy for critical correlations.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS12572593B2Methods and systems for building and/or using a graph data structure

Publication Date: 2026.03.10 XERO

AI summary

Methods comprising determining a dataset of entries, each comprising an entity identifier and associated application identifier(s) are provided. One method comprises determining a set of application pairs from the dataset; determining a correlation score for each application pair based; responsive to the correlation score exceeding a threshold correlation value: determining application identifiers of the application pair as being correlated; providing a correlation write request comprising the applications identifiers of the pair to an app recommender database.