Graph Pruning Connected Components Threshold

Solution Overview

Problem

The challenge is to efficiently process large graph datasets within the memory and disk capacity limitations of computing devices, as existing methods struggle to handle datasets exceeding typical machine capacities, leading to increased processing time and memory requirements without producing relevant information of interest.

Innovation Solution

A system that identifies connected components in a graph smaller than a minimum graph traversal threshold and prunes them, generating a second graph for processing, which reduces computation time and memory requirements by discarding unnecessary components, thereby facilitating efficient data extraction.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If automated compilation and formatting of data using scripts and UNIX utilities is performed, then the input graph dataset can be processed, but the processing time and memory requirements become excessively high when the dataset size exceeds typical machine capacities

Engineering Contradiction:

Improvedata processing capabilityVSAvoidprocessing time

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The patent extracts and removes connected components from the graph that are smaller than the minimum graph traversal threshold. This extraction eliminates unnecessary data portions that would consume processing resources without contributing to queries requiring a minimum traversal depth, thereby reducing processing time while maintaining productivity for relevant data.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent segments the graph data into connected components and evaluates each component independently against the minimum graph traversal threshold. This segmentation allows the system to process and filter components individually, reducing overall processing time by avoiding unnecessary traversal of small components that cannot satisfy query requirements.

Inventive Principle:

Principle #1Segmentation

2Productivity

If automated compilation and formatting of data using scripts and UNIX utilities is performed, then the input graph dataset can be processed, but the memory and disk space requirements exceed typical machine capacities

Engineering Contradiction:

Improvedata processing capabilityVSAvoidmemory and disk space

Core Design Contradiction:

ProductivityVSQuantity of substance

Solution Approach 1:

The patent extracts and removes connected components from the graph that are smaller than the minimum graph traversal threshold. This extraction eliminates unnecessary data portions that would consume processing resources without contributing to queries requiring a minimum traversal depth, thereby reducing processing time while maintaining productivity for relevant data.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent segments the graph data into connected components and evaluates each component independently against the minimum graph traversal threshold. This segmentation allows the system to process and filter components individually, reducing overall processing time by avoiding unnecessary traversal of small components that cannot satisfy query requirements.

Inventive Principle:

Principle #1Segmentation

3Loss of time

If the minimum graph traversal threshold is applied to prune small connected components, then processing time and memory usage are reduced, but the system must first identify and evaluate each component

Engineering Contradiction:

Improveprocessing timeVSAvoidsystem complexity

Core Design Contradiction:

Loss of timeVSDevice complexity

Solution Approach 1:

The system performs self-service by automatically identifying connected components, evaluating them against the minimum graph traversal threshold, and pruning unnecessary components without requiring external intervention. This automation reduces system complexity overhead while maintaining the benefits of threshold-based filtering.

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS9529833B2Graph pruning in hipergraph

Publication Date: 2016.12.27 GENESEE VALLEY INNOVATIONS LLC

AI summary

One embodiment of the present invention provides a system for graph pruning. During operation, the system identifies a connected component in a graph comprising one or more vertices. A respective vertex of the graph represent an element in a data set, an edge between two vertices represents a type and strength of relationship between the vertices. The system identifies a connected component to be smaller than a minimum graph traversal threshold associated with a query for the graph. This minimum graph traversal threshold indicates a minimum number of traversal steps needed for the query. The system then generates a second graph by pruning the connected component from the graph. This second graph is processed to extract information of interest from the data set.