Graph Data Structures Using Homomorphism to Reduce Redundancy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Large and interconnected data sets in graph databases lead to redundancy, making computational operations inefficient and challenging to track data provenance, with entity resolution being computationally intensive and error-prone.
Innovation Solution
A system generates transformed storage-efficient graph data structures by creating a reduced graph database through equivalence relations and connected-components, using graph homomorphisms to reduce redundancy and facilitate entity-specific queries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If graph databases store all relationship information to support complex queries, then query capability is improved, but data redundancy increases and computational efficiency deteriorates
Solution Approach 1:
The patent segments the graph database into multiple disconnected components, where each component contains a subset of the total nodes and edges. This segmentation allows computational operations to be performed on smaller, independent subsets rather than the entire graph, significantly improving computational efficiency while preserving the ability to answer queries within each component's scope.
Solution Approach 2:
The patent introduces a new dimension of decomposition by breaking down the graph database not just by size but by connectivity components. This dimensional change allows the system to maintain full query capability within each component while avoiding the computational overhead of processing the entire graph, effectively adding a structural organization layer that separates data storage from computational processing.
2Ease of operation
If multiple databases are created from parent databases to meet organizational needs, then data accessibility is improved, but data redundancy increases and lineage tracking becomes difficult
Solution Approach 1:
The patent creates simplified copies of the graph database structured as disconnected components, where each component is a self-contained subset that can be independently processed. These copies maintain the essential structural and relational information needed for queries while eliminating redundant data across multiple database instances, thus preserving data lineage without requiring complex tracking mechanisms.
Solution Approach 2:
The patent discards redundant information that exists across multiple database copies by identifying and removing duplicate nodes and edges. The essential information is recovered and preserved in the disconnected component structure, where each component contains only the unique data necessary for its specific queries, thereby maintaining data accessibility while eliminating redundancy and preserving lineage information.
3Measurement precision
If entity resolution is performed across multiple databases to identify the same physical entity, then data accuracy is improved, but computational complexity increases and errors increase
Solution Approach 1:
The patent segments the entity resolution problem by restricting comparisons to within disconnected components rather than across the entire graph database. Since nodes in different components are guaranteed to be unrelated, entity resolution only needs to be performed within each component's subset of nodes, dramatically reducing computational complexity while maintaining data accuracy through localized comparison operations.
4Speed
If graph databases store comprehensive relationship information, then relationship traversal efficiency is improved, but storage requirements increase and redundancy increases
Solution Approach 1:
The patent extracts only the essential nodes and edges that form each disconnected component, removing redundant relationship information that would otherwise be stored across multiple database instances. Each component contains precisely the relationship data needed for traversal within that component, maintaining traversal efficiency while minimizing storage requirements by eliminating duplicate relationships across the broader database system.
Data Source
AI summary
A system for generating a storage-efficient data structure representing a plurality of inter-related data tables and adapted for use in data processing. The system configured to receive input data and generate a graph database having edges and vertices. Each inter-related data table of a plurality of inter-related data tables is received as input data and defines a corresponding vertex of the vertices. The edges defining pairwise relationships between the vertices based on one or more common elements of a corresponding pair of data tables. A reduced graph database is generated, as well as its connected-components. A partition defined by an equivalence relation on the graph database is used to generate the reduced graph database. Each set of the plurality of sets of the partition is represented by a single vertex of the corresponding set. An output data structure indicative of the connected-components, or additionally or alternatively, the irreducible generators, is generated.