# US 10,762,293 B2 · Using parts-of-speech tagging and named entity recognition for spelling correction > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 10,762,293 B2 - Document ID: US10762293B2 - Status: Granted - Kind: B2 - Application number: 12/976,849 - Filed: 2010-12-22 - Published: 2012-06-28 - Granted: 2020-09-01 - Earliest claimed: 2010-12-22 - Assignees: Apple Inc. - Applicants: Brent D. Ramerth; Douglas R. Davidson; Jennifer Lauren Moore - Inventors: Brent D. Ramerth; Douglas R. Davidson; Jennifer Lauren Moore - Web page: https://patentgrep.com/patent/US10762293B2 ## CPC - G06F 40/232: Orthographic correction, e.g. spell checking or vowelisation - G06F 40/169: Annotation, e.g. comment data or footnotes - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) ## Abstract Techniques to automatically correct or complete text are disclosed. An entered text and a context data indicating a context in which the entered text is used are received. Examples of context data include additional words and/or a phrase or sentence in which the entered text occurs. A replacement candidate to replace the entered text is determined based on the entered text and the context data. ## Claims ### Claim 1 A method of automatically correcting or completing text, comprising: receiving entered text from a user and context data indicating a context in which the entered text is used, wherein the entered text is a word; determining without user intervention, based on the word and the context data, a replacement candidate to replace the word, the determining including: using the context data to assign to the replacement candidate a score indicating a degree of confidence that the replacement candidate should be suggested; determining a statistically expected part of speech of the word; and selecting the replacement candidate based at least in part on a determination that a part of speech of the replacement candidate matches the statistically expected part of speech of the word; and providing the selected replacement candidate to the user as a suggested correction. ### Claim 2 (depends on 1) The method of claim 1, wherein the replacement candidate comprises a more correct or complete word or phrase than the entered text. ### Claim 3 (depends on 1) The method of claim 1, wherein the context data comprises one or more words that occur in a same sentence as the entered text. ### Claim 4 (depends on 1) The method of claim 1, wherein the context data is used to determine one or more features of the entered text. ### Claim 5 (depends on 4) The method of claim 4, wherein the features include one or more of the following: an identification as a named entity, a prefix, a suffix, and capitalization. ### Claim 6 (depends on 1) The method of claim 1, further comprising receiving further context data as additional text is entered and using the further context data to update an evaluation of one or more replacement candidates. ### Claim 7 (depends on 1) The method of claim 1, wherein determining the replacement candidate includes using the context data to evaluate the replacement candidate. ### Claim 8 (depends on 1) The method of claim 1, wherein determining the replacement candidate further includes selecting the replacement candidate based at least in part on a determination that the score exceeds a selection threshold. ### Claim 9 (depends on 1) The method of claim 1, wherein determining the replacement candidate includes determining based at least in part on the context data that the replacement candidate is more likely correct than one or more alternative replacement candidates. ### Claim 10 (depends on 1) The method of claim 1, further comprising using a statistical language model, the word, and the context data to determine the statistically expected part of speech of the word. ### Claim 11 (depends on 10) The method of claim 10, further comprising generating the statistical model. ### Claim 12 (depends on 11) The method of claim 11, wherein generating the statistical model includes augmenting a commercially available annotated corpus with annotated content comprising one or more of blog entries, online comments, comments posted on online social networks, and other user generated online content. ### Claim 13 (depends on 1) The method of claim 1, further comprising: providing the selected replacement candidate to the user as a selectable suggested correction. ### Claim 14 (depends on 1) The method of claim 1, wherein determining the statistically expected part of speech of the word comprises determining a statistically expected lexical category of the word. ### Claim 15 A system, comprising: an input device configured to receive user inputs comprising entered text; and a processor coupled to the input device and configured to: receive entered text from a user entered using the input device and context data indicating a context in which the entered text is used, wherein the entered text is a word; determine without user intervention, based on the word and the context data, a replacement candidate to replace the word, the determining including: using the context data to assign to the replacement candidate a score indicating a degree of confidence that the replacement candidate should be suggested; determining a statistically expected part of speech of the word; and selecting the replacement candidate based at least in part on a determination that a part of speech of the replacement candidate matches the statistically expected part of speech of the word; and provide the selected replacement candidate to the user as a suggested correction. ### Claim 16 (depends on 15) The system of claim 15, wherein the context data comprises one or more words that occur in a same sentence as the entered text. ### Claim 17 (depends on 15) The system of claim 15, wherein the processor is configured to update an evaluation of the replacement candidate as additional context data is received. ### Claim 18 (depends on 15) The system of claim 15, wherein the processor is configured to provide the selected replacement candidate to the user as a selectable suggested correction. ### Claim 19 (depends on 15) The system of claim 15, wherein determining the statistically expected part of speech of the word comprises determining a statistically expected lexical category of the word. ### Claim 20 (depends on 15) The system of claim 15, wherein the replacement candidate comprises a more correct or complete word or phrase than the entered text. ### Claim 21 (depends on 15) The system of claim 15, wherein the context data is used to determine one or more features of the entered text. ### Claim 22 (depends on 21) The system of claim 21, wherein the features include one or more of the following: an identification as a named entity, a prefix, a suffix, and capitalization. ### Claim 23 (depends on 15) The system of claim 15, wherein determining the replacement candidate includes using the context data to evaluate the replacement candidate. ### Claim 24 (depends on 15) The system of claim 15, wherein determining the replacement candidate further includes selecting the replacement candidate based at least in part on a determination that the score exceeds a selection threshold. ### Claim 25 (depends on 15) The system of claim 15, wherein determining the replacement candidate includes determining based at least in part on the context data that the replacement candidate is more likely correct than one or more alternative replacement candidates. ### Claim 26 (depends on 15) The system of claim 15, wherein the processor is further configured to use a statistical language model, the word, and the context data to determine the statistically expected part of speech of the word. ### Claim 27 (depends on 26) The system of claim 26, wherein the processor is further configured to: generate the statistical model. ### Claim 28 (depends on 27) The system of claim 27, wherein generating the statistical model includes augmenting a commercially available annotated corpus with annotated content comprising one or more of Hog entries, online comments, comments posted on online social networks, and other user generated online content. ### Claim 29 A non-transitory computer readable storage medium storing one or more programs for execution by an electronic device, the one or more programs comprising instructions for automatically correcting or completing text, including: receiving entered text from a user and context data indicating a context in which the entered text is used, wherein the entered text is a word; determining without user intervention, based on the word and the context data, a replacement candidate to replace the word, the determining including: using the context data to assign to the replacement candidate a score indicating a degree of confidence that the replacement candidate should be suggested; determining a statistically expected part of speech of the word; and selecting the replacement candidate based at least in part on a determination that a part of speech of the replacement candidate matches the statistically expected part of speech of the word; and providing the selected replacement candidate to the user as a suggested correction. ### Claim 30 The non-transitory computer readable storage medium of claim the one or more programs further comprising instructions for providing the selected replacement candidate to the user as a selectable suggested correction. ### Claim 31 (depends on 29) The non-transitory computer readable storage medium of claim 29, wherein determining the statistically expected part of speech of the word comprises determining a statistically expected lexical category of the word. ### Claim 32 (depends on 29) The non-transitory computer readable storage medium of claim 29, wherein the replacement candidate comprises a more correct or complete word or phrase than the entered text. ### Claim 33 (depends on 29) The non-transitory computer readable storage medium of claim 29, wherein the context data comprises one or more words that occur in a same sentence as the entered text. ### Claim 34 (depends on 29) The non-transitory computer readable storage medium of claim 29, wherein the context data is used to determine one or more features of the entered text. ### Claim 35 (depends on 34) The non-transitory computer readable storage medium of claim 34, wherein the features include one or more of the following: an identification as a named entity, a prefix, a suffix, and capitalization. ### Claim 36 (depends on 29) The non-transitory computer readable storage medium of claim 29, wherein the one or more programs further comprise instructions for: receiving further context data as additional text is entered and using the further context data to update an evaluation of one or more replacement candidates. ### Claim 37 (depends on 29) The non-transitory computer readable storage medium of claim 29, wherein determining the replacement candidate includes using the context data to evaluate the replacement candidate. ### Claim 38 (depends on 29) The non-transitory computer readable storage medium of claim 29, wherein determining the replacement candidate further includes selecting the replacement candidate based at least in part on a determination that the score exceeds a selection threshold. ### Claim 39 (depends on 29) The non-transitory computer readable storage medium of claim 29, wherein determining the replacement candidate includes determining based at least in part on the context data that the replacement candidate is more likely correct than one or more alternative replacement candidates. ### Claim 40 (depends on 29) The non-transitory computer readable storage medium of claim 29, wherein the one or more programs further comprise instructions for: using a statistical language model, the word, and the context data to determine the statistically expected part of speech of the word. ### Claim 41 (depends on 40) The non-transitory computer readable storage medium of claim 40, wherein the one or more programs further comprise instructions for: generating statistical model. ### Claim 42 (depends on 41) The non-transitory computer readable storage medium of claim 41, wherein generating the statistical model includes augmenting a commercially available annotated corpus with annotated content comprising one or more of blog entries, online comments, comments posted on online social networks, and other user generated online content.