Dynamic Term Synonym Generation for Search Indexing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Search engines face challenges in returning accurate results when users input only a portion of a term or improperly format search terms, especially for non-word searches like part numbers, due to the vast amount of information available online.

Innovation Solution

Dynamically generating and indexing synonyms for a term by applying rules that include only parts of the term or modify it by adding or removing characters, to improve search accuracy and account for user errors.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the search engine indexes only the exact term provided, then the indexing process is simple and fast, but the search accuracy decreases when users input partial or incorrectly formatted terms

Engineering Contradiction:
Improvesearch accuracyVSAvoidindexing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system performs preliminary generation of synonym terms during the indexing phase. Before the actual search occurs, the index is pre-populated with multiple variant forms of each term (including partial matches, formatted variations, and common misspellings). This preliminary action ensures that when users search with imperfect queries, the results are already optimized for accuracy without adding complexity to the search-time processing.

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If the search engine indexes multiple synonyms and variations for each term, then search accuracy improves for partial or incorrect inputs, but the indexing time and computational resources increase

Engineering Contradiction:
Improvesearch accuracyVSAvoidindexing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The synonym generation process is segmented into distinct, manageable components. The system divides term variations into categories such as partial matches, formatted variations, and common errors. Each segment is processed independently using targeted rules, allowing the indexing system to efficiently generate only the most relevant synonyms rather than exhaustively processing all possible variations, thus reducing overall indexing time while maintaining search accuracy.

Inventive Principle:
Principle #1Segmentation

3Adaptability or versatility

If the system generates synonyms by including only parts of the term, then it helps users who know only a portion of the identifier, but it increases the number of terms to be indexed

Engineering Contradiction:
Improvesearch flexibilityVSAvoidnumber of indexed terms
Core Design Contradiction:
Adaptability or versatilityVSQuantity of substance

Solution Approach 1:

The system applies local quality by selectively generating partial-term synonyms only for specific positions within the original term. Rather than creating all possible substrings, the system identifies and generates partial matches at locally significant positions (such as beginning, middle, or end segments that are most likely to be user-known). This targeted approach enhances search flexibility for users who know partial identifiers while controlling the overall quantity of indexed terms.

Inventive Principle:
Principle #3Local quality

Data Source

PatentUS8065283B2Term synonym generation
Publication Date: 2011.11.22 GSCOMPARE LLC
  • US8065283B2 patent drawing
  • US8065283B2 patent drawing
  • US8065283B2 patent drawing

AI summary

Synonyms for a term to be indexed are dynamically generated by applying one or more rules (e.g., splitting, deletion or addition of characters, and concatenation of portions) to the term, each synonym generated either including only a portion and not all of the term or differing from the term by at least one additional character in a position between a first character and a last character (i.e., not at either end of the term). The term and some or all of the synonyms are then indexed for subsequent searching.