System and method for disambiguating data to improve analysis of electronic content
US12045561B2 · kind B2 · utility
Assignee
Inventors
Key dates
| Filing date | Nov 28, 2022 |
| Grant date | Jul 23, 2024 |
| Priority date | — |
| Expiry date | Nov 28, 2042 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06F40/253
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
Systems and methods are disclosed for using natural language processing to generate sound-alikes and look-alikes for potentially relevant terms that could be mis-transcribed in a transcript. The disclosed framework combines the advantages of several measures aimed at tackling a wide range of transcription errors including but not limited to word boundary errors, phonetic confusion of words, spelling mistakes, grammatical errors, character drops, etc. In some examples, morphological similarity and phonetic similarity are used to address the word boundary errors and phonetic confusion of words. In some examples, a spell checker, word formation, and look-alike sound-alike generator are used to address errors such as phonetic confusion, spelling mistakes and grammatical errors. The generated sound-alikes can be ranked, which enables a flexible application of the generated phrases.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.