Google Inc.

Dan Popovici

Dimitra Papachristou

Robert Spalek

Kendra Jay Carattini

Terry Yang-Hoe Koo

Lina C Martensson

Jeremy D Hoffman

Xin Zheng

Juraj Gottweis

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for obtaining one or more first search results that were generated for a search query; determining a score associated with the first search results; revising the search query using a query revision rule; obtaining one or more second search results that were generated for the revised search query; determining a score associated with the second search results; and evaluating the query revision rule by comparing the score associated with the first search results with the score associated with the second search results.

Synonym identification based on search quality

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for query synonym expansion. One method includes receiving a query including a first compound term, and in response to receiving the query, performing the following operations before search results responsive to the query are identified: generating one or more splits of the first compound term, wherein each split divides the compound term into two or more subterms, assigning a score to each subterm of each split, determining an overall score for each split from the scores for the subterms of the split, selecting one or more of the one or more splits according to the overall score for each split, and augmenting the query with the subterms of each selected split.

Synonym generation using online decompounding and transitivity

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for click or skip evaluation of proximity rules. In one aspect, a method includes accessing query data that identifies, for a search query, a particular query term and other query terms included in the search query, search results, and a particular search result selected by a user. The method further includes determining, using the query data, that, (i) in text associated with the particular search result, the particular query term is separated from the other query terms by a minimum number of terms, and (ii) in text associated with a search result that was ranked higher than the particular search result, the particular query term is not separated from the other query terms by the minimum number of terms, then incrementing a click count for a query term proximity rule corresponding to the particular query term.

Click or skip evaluation of proximity rules

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for click or skip evaluation of optionalization rules. In one aspect, a method includes accessing query log data that identifies, for an initial search query, one or more revised search queries that were identified by designating a query term of the initial search query as optional according to one or more optionalization rules, and a particular search result selected by the user. A determination is made, using the query log data, that the particular search result corresponds to the one or more revised search queries that were identified by designating the query term of the initial search query as optional according to one or more optionalization rules. In response to this determination, a click count is incremented for the one or more optionalization rules that corresponds to the particular search result.

Click or skip evaluation of query term optionalization rule

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for click or skip evaluation of synonyms. In one aspect, a method includes accessing query log data that identifies, for an initial search query, a revised search query that includes one or more synonyms, and search results provided to a user. An indication that the user has selected a particular search result is received. A determination is made that the particular search result includes a particular synonym for a query term that occurs in the revised search query, and not (i) the query term or (ii) any other synonym for the query term that occurs in the revised search query. A crucial click count is assigned to a synonym rule that corresponds to the particular synonym.

Click or skip evaluation of synonym rules

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for removing substitution rules. According to one implementation, a method includes identifying a plurality of revised search queries associated with a substitution rule, wherein the substitution rule is used to generate the revised search queries from each of a plurality of original search queries to include a substitute query term for an original query term included in the original search queries; for each of the revised search queries, accessing query log data that identifies (i) a set of search results generated by the respective revised search query, and (ii) a particular, user-selected search result in the set of search results; determining a quality metric for the substitution rule using the query log data; determining that the quality metric does not satisfy a threshold; and in response to determining that the quality metric does not satisfy the threshold, removing the substitution rule.

Removing substitution rules based on user interactions

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for evaluating a query term reordering rule. In one aspect, a method includes accessing query log data that identifies, for an initial search query, two or more query terms that are included in a particular order in the initial search query, search results that were generated using the initial search query, and a particular search result selected by the user. A determination is made, using the query log data, that the particular search result includes the two or more query terms in a different order than the particular order in which the two or more query terms are ordered in the initial search query. In response to this determination, a click count is incremented for a query term reordering rule associated with the two or more query terms.

Click or skip evaluation of reordering rules

Methods, systems, and apparatus, including computer programs are encoded on a computer storage medium, for fake skip evaluation of synonyms. In one aspect, a method includes determining, using query log data, that a particular search result selected by a user includes a query term included in an initial search query and a particular synonym that was generated for the query term using a particular synonym rule. The particular search result is selected by the user from among search results that were generated using an initial search query and one or more revised search queries that include the particular synonym. The method further includes determining, using the query log data, that a first search result is ranked above the particular search result, and includes the particular synonym for the query term. In response to these determinations, a fake skip count is incremented for the synonym rule that corresponds to the particular synonym.

Fake skip evaluation of synonym rules

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for adding synonym rules based on historic data. According to one implementation, a method includes receiving an indication that a synonym rule that was previously added to a collection of synonym rules is no longer being added to the collection, and receiving historic usage data relating to the synonym rule. The method also includes determining, using the historic usage data relating to the synonym rule, whether to add the synonym rule to the collection, and, in response to determining to add the synonym rule, adding the synonym rule to the collection.

Adding synonym rules based on historic data

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for removing substitution rules. According to one implementation, a method includes identifying a revised search query that was revised to include a substitute term of a query term; identifying search results that were generated using the revised search query, wherein each search result references a resource; determining, by one or more computers, that none of the resources referenced by a subset of the search results include the substitute term of the query term; and in response to determining that none of the resources referenced by the subset of search results include the substitute term of the query term, incrementing a no-match score for the substitute term.

Removing substitution rules

Online de-compounding of query terms

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for query synonym expansion. One method includes receiving a query including a first compound term, and in response to receiving the query, performing the following operations: generating one or more splits of the first compound term, assigning a score to each subterm of each split, determining an overall score for each split from the scores for the subterms of the split, selecting a first split from the one or more splits according to the overall score for each split, and augmenting the query with a first synonym phrase that is a synonym of a first subterm of the first split.

Growing community of inventors

Cupertino, CA, United States of America

Dan Popovici