# US 8,572,081 B1 · Identifying non-compositional compounds > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 8,572,081 B1 - Document ID: US8572081B1 - Status: Granted - Kind: B1 - Application number: 13/361,565 - Filed: 2012-01-30 - Granted: 2013-10-29 - Earliest claimed: 2009-03-12 - Assignees: Google Inc. - Applicants: Stewart Yang; Fang Liu; Pei Cao - Inventors: Stewart Yang; Fang Liu; Pei Cao - Web page: https://patentgrep.com/patent/US8572081B1 ## CPC - G06F 40/289: Phrasal analysis, e.g. finite state techniques or chunking - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) ## Abstract Methods, systems, and apparatus, including computer programs encoded on computer storage media, for identifying non-compositional compounds. In one aspect, a method includes the actions of receiving a collection of phrases, each phrase including two or more words; for each phrase, determining if the phrase is a non-compositional compound, a non-compositional compound being a phrase of two or more words where the words composing the phrase have different meanings in a compound than their conventional meanings individual, the determining including: identifying a similar term for a term of the phrase, substituting the similar term for the term of the phrase to generate a substitute phrase, calculating a similarity between the phrase and the substitute phrase, and identifying the phrase as a non-compositional compound when the calculated similarity is less than a specified threshold value. ## Claims ### Claim 1 A method comprising: receiving a collection of phrases from a collection of text, each phrase including two or more words; and for each phrase, determining if the phrase is a non-compositional compound, a non-compositional compound being a phrase of two or more words where the words composing the phrase have different meanings in a compound than their conventional meanings individually, the determining including: identifying a similar word for a word of the phrase, substituting the similar word for the word of the phrase to generate a substitute phrase, calculating a similarity between the phrase and the substitute phrase, including performing a similarity measure based on a context of the phrase and the substitute phrase, respectively, in the collection of text, and identifying the phrase as a non-compositional compound based on a comparison of the calculated similarity with a specified threshold. ### Claim 2 (depends on 1) The method of claim 1, wherein the similarity identifies a category similarity. ### Claim 3 (depends on 1) The method of claim 1, further comprising annotating phrases identified as non-compositional compounds. ### Claim 4 (depends on 1) The method of claim 1, further comprising: receiving a search query; determining whether the search query includes a non-compositional compound; and modifying a search for resources responsive to the search query when the search query includes a non-compositional compound; and presenting search results in response to the modified search. ### Claim 5 (depends on 4) The method of claim 4, wherein the modifying includes searching for the non-compositional compound and not constituent terms of the non-compositional phrase. ### Claim 6 (depends on 1) The method of claim 1, further comprising: receiving a search query; determining whether the search query includes a non-compositional compound; using the non-compositional compound to identify one or more advertisements; and presenting search results responsive to the search query and the one or more advertisements. ### Claim 7 A method comprising: receiving a collection of phrases from a collection of text, each phrase including two or more words; for each phrase, determining if the phrase is a non-compositional compound, the determining including: identifying a similar word for a word of the phrase, substituting the similar word for the word of the phrase to generate a substitute phrase, calculating a similarity between the phrase and the substitute phrase, and identifying the phrase as a non-compositional compound based on a comparison of the calculated similarity with a specified threshold; and annotating each phrase that is identified as a non-compositional compound. ### Claim 8 (depends on 7) The method of claim 7, wherein a non-compositional compound is a phrase of two or more words where the words composing the phrase have different meanings in a compound than their conventional meanings individually. ### Claim 9 (depends on 7) The method of claim 7, wherein identifying similar words to a non-compositional compound word includes performing a similarity measure based on the context of a candidate word and the context of other words from a collection of text. ### Claim 10 (depends on 7) The method of claim 7, wherein calculating similarity between candidate non-compositional compound and substitute phrase includes performing a similarity measure based on the context of the candidate non-compositional compound and the context of the substitute phrase from a collection of text. ### Claim 11 (depends on 7) The method of claim 7, wherein receiving phrases comprises: receiving a collection of text; and extracting phrases from the collection of text. ### Claim 12 (depends on 7) The method of claim 7, wherein the similarity is a vertical similarity. ### Claim 13 (depends on 12) The method of claim 12, wherein the vertical similarity refers to a category similarity. ### Claim 14 (depends on 7) The method of claim 7, wherein the similarity is a horizontal similarity. ### Claim 15 (depends on 14) The method of claim 14, wherein the horizontal similarity is based on circumstances in which the phrase and the substitute phrase occur together. ### Claim 16 (depends on 7) The method of claim 7, further comprising: receiving a search query; determining whether the search query includes a non-compositional compound; and modifying a search for resources responsive to the search query when the search query includes a non-compositional compound; and presenting search results responsive to the search query. ### Claim 17 (depends on 16) The method of claim 16, wherein the modifying includes searching for the non-compositional compound and not constituent terms of the non-compositional phrase. ### Claim 18 (depends on 7) The method of claim 7, further comprising: receiving a search query; determining whether the search query includes a non-compositional compound; using the non-compositional compound to identify one or more advertisements; and presenting search results responsive to the search query and the one or more advertisements. ### Claim 19 A method comprising: receiving a collection of phrases from a collection of text, each phrase including two or more words; for each phrase, determining if the phrase is a non-compositional compound, the determining including: identifying a similar word for a word of the phrase, substituting the similar word for the word of the phrase to generate a substitute phrase, calculating a similarity between the phrase and the substitute phrase, including performing a similarity measure based on a context of the phrase and the substitute phrase, respectively, in the collection of text, and identifying the phrase as a non-compositional compound based on a comparison of the calculated similarity with a specified threshold; and annotating each phrase that is identified as a non-compositional compound.