Phonetic Search Term Ranking for Misspellings and Homophones

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional search systems struggle to provide high-quality search term suggestions for misspelled or homophone-based queries due to limitations in phonetic algorithms and edit distance methods, which often fail to account for similar sounds or frequencies of terms.

Innovation Solution

A search system utilizing multiple phonetic encoder modules to generate candidate terms based on sound codes and frequency of appearance in the corpus, ranking these terms by sound distance and frequency, and maintaining a separate sound index for efficient retrieval of candidate terms.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If edit distance algorithms are used to provide search term suggestions, then misspelled terms can be corrected, but terms with similar sounds or pronunciations are not effectively addressed

Engineering Contradiction:

Improvespell checking accuracyVSAvoidhandling of homophones

Core Design Contradiction:

Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent introduces phonetic encoders as intermediary components that convert search terms into phonetic representations. These encoders act as mediators between the original search term and the suggestion system, enabling the detection of homophones and similar-sounding terms that edit distance algorithms miss. The phonetic encoder transforms textual input into a phonetic space where sound-based similarities can be effectively measured and matched.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Adaptability or versatility

If larger edit distances are used to find more suggestions, then more alternative terms can be found, but search performance deteriorates

Engineering Contradiction:

Improvenumber of suggestionsVSAvoidsearch performance

Core Design Contradiction:

Adaptability or versatilityVSProductivity

Solution Approach 1:

The patent replaces the mechanical edit distance calculation system with a phonetic encoding system. Instead of computationally intensive character-by-character comparison with large edit distances, the system uses phonetic encoders to transform terms into phonetic representations, enabling faster and more accurate identification of sound-based similarities without the performance penalty of large edit distance searches.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Reliability

If conventional spell checkers are used, then correctly spelled words can be suggested, but misspelled or homophone search terms are not effectively handled

Engineering Contradiction:

Improvespell checkingVSAvoidhandling of search variations

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent adds a phonetic dimension to the traditional spell-checking approach. Instead of operating solely in the textual/spelling dimension, the system transforms search terms into a phonetic dimension using encoders. This additional dimension enables the system to detect and suggest corrections for homophones and sound-based variations that conventional spell checkers cannot handle, while maintaining the reliability of traditional spelling correction.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

Data Source

PatentUS12613891B2System and method for recommendation of terms, including recommendation of search terms in a search system

Publication Date: 2026.04.28 OPEN TEXT SA ULC

AI summary

Embodiments of systems and methods for providing search term suggestions in a search system are disclosed. Embodiments as disclosed may utilize the sound of an original search term to locate candidate terms based on the sound of the candidate terms and the frequency of appearance of the candidate terms in the corpus of documents being searched. A set of search term suggestions can then be determined from the candidate terms and returned to the user as search term suggestions for the original search term.