Graph Node Repositioning via Influencing Parameters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing computational linguistic analysis techniques for large document collections often result in graphs that are misaligned due to factors like text length variations and jargon usage, leading to inefficient resource utilization and less relevant results, as certain objects are relegated to marginal positions due to poor descriptions.
Innovation Solution
The method involves selecting influencing parameters to modify or supplement the graph, allowing nodes to be reassessed based on these parameters, which can adjust edge weights and connectivity, thereby improving the representation of objects and enhancing computational efficiency and result relevance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If graph nodes are positioned based on term frequency in plain text, then the graph can be constructed from available text data, but text length variations and jargon usage cause certain objects to be relegated to marginal positions, undermining the explanatory power of the graph
Solution Approach 1:
The patent changes the parameter used for graph construction from raw term frequency to a normalized metric that accounts for text length variations. Specifically, it uses the ratio of term frequency to document length (or similar normalization), which adjusts the weight of terms based on the size of the source text. This prevents short documents with dense jargon from unfairly dominating the graph topology and ensures that objects are positioned according to their actual significance across the corpus rather than artifacts of text length.
2Quantity of substance
If the graph includes all documents in the collection, then comprehensive coverage is achieved, but computational resources are inefficiently utilized and results become less relevant
Solution Approach 1:
The patent extracts and removes less relevant documents or nodes from the graph based on threshold criteria. It identifies documents that contribute minimally to the overall graph structure or analysis goals and excludes them from the final graph representation. This extraction process maintains the essential information while reducing computational overhead, allowing the system to focus resources on the most significant portions of the document collection.
Solution Approach 2:
The patent segments the document collection into different tiers or groups based on their relevance or contribution to the graph. Instead of treating all documents uniformly, it creates hierarchical or modular graph structures where core documents form the central network and peripheral documents are either excluded or given lesser weight. This segmentation enables more efficient computation by allowing selective processing of different document subsets.
3Ease of manufacture
If edge weights are determined solely by term frequency, then the graph construction is simple, but the representation becomes misleading due to poor text descriptions and uncommon words
Solution Approach 1:
The patent combines multiple factors into a composite edge weight calculation rather than relying on a single metric. It integrates term frequency with additional signals such as text length normalization, document relevance scores, or contextual similarity measures. This composite approach creates a more robust edge weight that reflects both the presence of terms and the quality of their contextual usage, thereby improving the accuracy of object similarity representation while maintaining reasonable computational complexity.
Data Source
AI summary
Provided is a process including: obtaining a graph comprising nodes and edges, each of the edges having a value indicating an amount of similarity between objects corresponding to the two linked nodes; selecting a parameter for influencing the graph; assessing each of the nodes based on the selected influencing parameter, wherein assessing comprises, with respect to each adjacent node in the graph sharing an edge with the node: determining the value indicating the amount of similarity between the object corresponding to the node and the object corresponding to the adjacent node; and determining a score related to the edge shared with the node, the score determined based on the similarity-amount value and a value of the selected influencing parameter for the node, such that edges are removed, weakened, added, or strengthened; and preparing, based on the graph, instructions to display at least part of the graph.