Graph Database Relevance Scoring via Pathway Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing news delivery systems often fail to identify relevant news articles that impact companies or portfolios without explicit mentions, leading to missed connections between news stories and entities of interest.
Innovation Solution
A graph-based computer-implemented method that quantifies the relevance between entities in a graph database by identifying and analyzing pathways connecting nodes, using predetermined criteria and weighting factors to compute pathway scores and generate visual representations of relationships.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If news articles are selected based on explicit mentions of companies in user portfolios, then the selection process is simple and fast, but relevant news that impacts companies without explicit mentions is missed
Solution Approach 1:
The patent introduces intermediate entities (suppliers, customers, competitors, industry groups) as mediators to connect news articles to companies indirectly. These intermediaries serve as bridges that reveal hidden relationships, allowing the system to identify relevant news through multi-hop pathways in the knowledge graph rather than relying solely on explicit mentions.
Solution Approach 2:
The patent transitions from a single-dimension approach (explicit text mentions) to a multi-dimensional approach by incorporating various relationship types (supplier, customer, competitor, industry group memberships) and multiple pathway lengths. This dimensional expansion enables the system to capture indirect relationships that span across different entity types and connection depths.
2Measurement precision
If multiple pathways and criteria are used to identify hidden connections, then news relevance accuracy improves, but computational complexity and processing time increase
Solution Approach 1:
The patent employs parameter changes by introducing weighted scoring mechanisms that evaluate multiple pathways with different importance weights. The system dynamically adjusts pathway scores based on relationship types, pathway lengths, and entity attributes, allowing precise relevance measurement while managing computational complexity through parameter-based prioritization.
Solution Approach 2:
The patent implements partial action by limiting the search to pathways within a specified maximum length and selecting only the top-scoring pathways rather than analyzing all possible connections. This selective approach achieves sufficient precision for practical applications without the prohibitive computational cost of exhaustive analysis.
3Adaptability or versatility
If indirect relationships through multiple intermediaries are tracked, then hidden connections are discovered, but the number of pathways to analyze increases exponentially
Solution Approach 1:
The patent applies preliminary action by pre-computing and storing relationship pathways in a knowledge graph structure before news analysis is needed. The system maintains pre-established connections between entities across multiple relationship types, enabling rapid querying and pathway identification during actual news relevance assessment without performing exhaustive searches at runtime.
Data Source
AI summary
In accordance with one aspect, the present disclosure is directed to quantifying the relevance between two entities in a graph database. One way of doing this is determining the number of pathways and the length of the pathways within the graph database that connect the two entities, wherein each entity is represented as a node. In some implementations, the present disclosure identifies and quantifies the connection between a news document (e.g., a news article) and a company or other entity. For example, the connection between an online news article and a company's underlying equity instrument can be evaluated.