Extended Synonym Generation for Search Precision

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional search systems face challenges in accurately searching unstructured data, often resulting in numerous irrelevant hits when using synonyms, as they fail to effectively leverage the hierarchical relationships within structured data.

Innovation Solution

A search support system that includes a synonym extraction section and an extended synonym generation section, which combine subject names from structured data with their synonyms to create extended synonyms, enhancing search accuracy by utilizing hierarchical relationships.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If synonyms are used to expand search coverage, then the quantity of search results increases, but the precision of search results deteriorates due to many non-associated data hits

Engineering Contradiction:
Improvequantity of search resultsVSAvoidprecision of search results
Core Design Contradiction:
Quantity of substanceVSMeasurement precision

Solution Approach 1:

The patent segments the synonym expansion process into two distinct stages: first extracting basic synonyms from teacher data, then generating extended synonyms by combining these basic synonyms with hierarchical subject names from structured data. This segmentation allows the system to maintain precision while expanding search coverage, as each stage builds upon the previous one with increasing specificity.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a hierarchical dimension to synonym generation by integrating structured data with predefined hierarchies (such as classification taxonomies). Instead of treating all synonyms equally, the system organizes them into hierarchical levels, allowing searches to traverse multiple dimensions of meaning and context, thereby improving precision while maintaining quantity.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Adaptability or versatility

If traditional synonym extraction from teacher data is used, then search coverage is improved, but search accuracy deteriorates due to lack of contextual hierarchy

Engineering Contradiction:
Improvesearch coverageVSAvoidsearch accuracy
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent merges two previously separate approaches: unstructured synonym extraction from teacher data and structured hierarchical classification. By combining basic synonyms with hierarchical subject names to create extended synonyms, the system simultaneously achieves broad search coverage and high search accuracy, as the extended synonyms capture both semantic equivalence and contextual hierarchy.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The extended synonym functions as a composite structure combining elements from both unstructured teacher data and structured hierarchical systems. This composite approach allows the synonym to carry multiple layers of meaning and context, improving search accuracy while maintaining the versatility needed for comprehensive search coverage.

Inventive Principle:
Principle #40Composite materials

Data Source

PatentUS11100149B2Search support system, search support method, and search support program
Publication Date: 2021.08.24 HITACHI LTD
  • US11100149B2 patent drawing
  • US11100149B2 patent drawing
  • US11100149B2 patent drawing

AI summary

A search support system supporting a search of unstructured data associated with a subject name that is a name of a subject indicated by a predetermined structure in structured data which contains a plurality of structures having a hierarchical relationship, is configured to include: a synonym search section that extracts a set of synonyms from teacher data which is unstructured data for detecting synonyms of the subject name; and an extended synonym generation section that generates an extended synonym which is a phrase obtained by combining a first subject name of a first structure in the structured data or a first synonym which is a synonym of the first subject name and which is extracted by the synonym search section with one or more second subject names indicated by one or more second structures higher in order than the first structure.