Evaluation of substitution contexts
View Patent ↗Methods, systems, and apparatus, including computer programs encoded on computer storage media, for evaluating substitute terms. One of the methods includes receiving a query and identifying an original term occurring in the query. One or more substitution contexts are determined for the original term, wherein a substitution context includes one or more context terms and an indication of a position in the query of the original term and the one or more context terms. A score is determined for each substitution context of the one or more substitution contexts. One or more substitution contexts are selected based on the score of each substitution context. One or more substitute terms for the original term in the context of the one or more selected substitution contexts are determined.
1. A computer-implemented method comprising:
receiving a query;
identifying an original term occurring in the query;
determining a plurality of substitution contexts for the original term occurring in the query, wherein a substitution context specifies a context in which the original term in the query is replaced by a substitute term, and wherein a substitution context includes one or more context terms and an indication of a position in the query of the original term relative to the one or more context terms;
generating a context hierarchy of the plurality of substitution contexts, wherein any conditions of a parent context in the context hierarchy also apply to each of one or more child contexts of the parent context in the context hierarchy, and wherein each child context in the context hierarchy includes all of the terms of a parent context of the child context as well as an additional term that does not occur in the parent context of the child context;
determining a score for each substitution context of the plurality of substitution contexts including comparing a particular substitution context to its parent context in the context hierarchy;
ranking the plurality of substitution contexts based on the score of each substitution context; and
determining one or more substitute terms for the original term occurring in the query in a context specified by a highest-ranking substitution context of the plurality of substitution contexts.
2. The method of claim 1 , further comprising:
classifying the plurality of substitution contexts into a first category or a second category based on the respective score of each substitution context; and
ranking substitution contexts in the first category.
3. The method of claim 1 , wherein determining a score for a substitution context comprises:
comparing a particular substitution context to the general context; and
computing a combined score using the comparison of the particular substitution context to the general context and the comparison of the particular substitution context to the general context.
4. The method of claim 1 , where the score represents a measure of how much meaning the substitution context adds to the original term compared to the parent substitution context.
5. The method of claim 1 , wherein determining a score for a substitution context comprises:
generating first frequencies of occurrence for alternate terms that occur in the substitution context in a textual collection;
generating second frequencies of occurrence for alternate terms that occur in the parent substitution context in the textual collection; and
comparing the first frequencies of occurrence and the second frequencies of occurrence.
6. The method of claim 1 , wherein computing a score for a substitution context having multiple parent contexts in the context hierarchy comprises:
determining a first score for a first parent substitution context of the substitution context;
determining a second score for a second parent substitution context of the substitution context; and
selecting the minimum of the first score or the second score as the score for the substitution context.
7. The method of claim 1 , comprising:
computing a score between two of the plurality of substitution contexts, wherein the score represents how closely the substitution contexts agree;
determining, based on the score, that the substitution contexts do not agree; and
in response to determining that the substitution contexts do not agree, decreasing a weight given to substitute terms for the original term generated using the substitution contexts that do not agree.
8. The method of claim 1 , comprising:
computing a score between two of the plurality of substitution contexts, wherein the score represents how closely the substitution contexts agree;
determining, based on the score, that the substitution contexts agree; and
in response to determining that the substitution contexts agree, increasing a weight given to substitute terms for the original term generated using the substitution contexts that agree.
9. A system comprising:
one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising:
receiving a query;
identifying an original term occurring in the query;
determining a plurality of substitution contexts for the original term occurring in the query, wherein a substitution context specifies a context in which the original term in the query is replaced by a substitute term, and wherein a substitution context includes one or more context terms and an indication of a position in the query of the original term relative to the one or more context terms;
generating a context hierarchy of the plurality of substitution contexts, wherein any conditions of a parent context in the context hierarchy also apply to each of one or more child contexts of the parent context in the context hierarchy, and wherein each child context in the context hierarchy includes all of the terms of a parent context of the child context as well as an additional term that does not occur in the parent context of the child context;
determining a score for each substitution context of the plurality of substitution contexts including comparing a particular substitution context to its parent context in the context hierarchy;
ranking the plurality of substitution contexts based on the score of each substitution context; and
determining one or more substitute terms for the original term occurring in the query in a context specified by a highest-ranking substitution context of the plurality of substitution contexts.
10. The system of claim 9 , wherein the operations further comprise:
classifying the plurality of substitution contexts into a first category or a second category based on the respective score of each substitution context; and
ranking substitution contexts in the first category.
11. The system of claim 9 , wherein determining a score for a substitution context comprises:
comparing a particular substitution context to the general context; and
computing a combined score using the comparison of the particular substitution context to the general context and the comparison of the particular substitution context to the general context.
12. The system of claim 9 , where the score represents a measure of how much meaning the substitution context adds to the original term compared to the parent substitution context.
13. The system of claim 9 , wherein determining a score for a substitution context comprises:
generating first frequencies of occurrence for alternate terms that occur in the substitution context in a textual collection;
generating second frequencies of occurrence for alternate terms that occur in the parent substitution context in the textual collection; and
comparing the first frequencies of occurrence and the second frequencies of occurrence.
14. The system of claim 9 , wherein computing a score for a substitution context having multiple parent contexts in the context hierarchy comprises:
determining a first score for a first parent substitution context of the substitution context;
determining a second score for a second parent substitution context of the substitution context; and
selecting the minimum of the first score or the second score.
15. The system of claim 9 , wherein the operations comprise:
computing a score between two of the plurality of substitution contexts, wherein the score represents how closely the substitution contexts agree;
determining, based on the score, that the substitution contexts do not agree; and
in response to determining that the substitution contexts do not agree, decreasing a weight given to substitute terms for the original term generated using the substitution contexts that do not agree.
16. The system of claim 9 , wherein the operations comprise:
computing a score between two of the plurality of substitution contexts, wherein the score represents how closely the substitution contexts agree;
determining, based on the score, that the substitution contexts agree; and
in response to determining that the substitution contexts agree, increasing a weight given to substitute terms for the original term generated using the substitution contexts that agree.
17. A computer program product, encoded on one or more non-transitory computer storage media, comprising instructions that when executed by one or more computers cause the one or more computers to perform operations comprising:
receiving a query;
identifying an original term occurring in the query;
determining a plurality of substitution contexts for the original term occurring in the query, wherein a substitution context specifies a context in which the original term in the query is replaced by a substitute term, and wherein a substitution context includes one or more context terms and an indication of a position in the query of the original term relative to the one or more context terms;
generating a context hierarchy of the plurality of substitution contexts, wherein any conditions of a parent context in the context hierarchy also apply to each of one or more child contexts of the parent context in the context hierarchy, and wherein each child context in the context hierarchy includes all of the terms of a parent context of the child context as well as an additional term that does not occur in the parent context of the child context;
determining a score for each substitution context of the plurality of substitution contexts including comparing a particular substitution context to its parent context in the context hierarchy;
ranking the plurality of substitution contexts based on the score of each substitution context; and
determining one or more substitute terms for the original term occurring in the query in a context specified by a highest-ranking substitution context of the plurality of substitution contexts.
18. The computer program product of claim 17 , wherein determining a score for a substitution context comprises:
comparing a particular substitution context to the general context; and
computing a combined score using the comparison of the particular substitution context to the general context and the comparison of the particular substitution context to the general context.
19. The computer program product of claim 17 , where the score represents a measure of how much meaning the substitution context adds to the original term compared to the parent substitution context.
20. The computer program product of claim 17 , wherein determining a score for a substitution context comprises:
generating first frequencies of occurrence for alternate terms that occur in the substitution context in a textual collection;
generating second frequencies of occurrence for alternate terms that occur in the parent substitution context in the textual collection; and
comparing the first frequencies of occurrence and the second frequencies of occurrence.
21. The computer program product of claim 17 , wherein computing a score for a substitution context having multiple parent contexts in the context hierarchy comprises:
determining a first score for a first parent substitution context of the substitution context;
determining a second score for a second parent substitution context of the substitution context; and
selecting the minimum of the first score or the second score.
22. The computer program product of claim 17 , wherein the operations comprise:
computing a score between two of the plurality of substitution contexts, wherein the score represents how closely the substitution contexts agree;
determining, based on the score, that the substitution contexts do not agree; and
in response to determining that the substitution contexts do not agree, decreasing a weight given to substitute terms for the original term generated using the substitution contexts that do not agree.
23. The computer program product of claim 17 , wherein the operations comprise:
computing a score between two of the plurality of substitution contexts, wherein the score represents how closely the substitution contexts agree;
determining, based on the score, that the substitution contexts agree; and
in response to determining that the substitution contexts agree, increasing a weight given to substitute terms for the original term generated using the substitution contexts that agree.