Recommendations based on branding

Abstract

A method and a system for providing recommendations based on branding are disclosed. In example embodiments, an index comprising predetermined brand relationships is maintained. Each predetermined brand relationship comprises a first brand, a second brand, and a recommendation score between the first brand and the second brand. A corpus containing a plurality of user queries is also maintained. A seed set of brands corresponding to a category in the index is expanded by accessing the corpus containing the plurality of user queries, evaluating user queries of the plurality of user queries that contain a disjunction of brand terms, and identifying a new brand to add to the seed set based on the evaluating.

1 . (canceled) 
     
     
         2 . A method, comprising:
 identifying, using a database of user queries, a plurality of brands included in the user queries;   for ones of the plurality of brands, identifying one or more non-brand terms associated with respective ones of the brands;   determining, for pairs of brands of the plurality of brands, respective recommendation scores that are based on the associated non-brand terms for corresponding pairs of brands, wherein a given recommendation score is indicative of a strength of relationship between a respective pair of brands;   in response to an indication of a user activity performed on a computing device by a user, identifying a preferred brand for the user; and   selecting, based on the respective recommendation scores, one or more different brands as a recommendation to the user.   
     
     
         3 . The method of  claim 2 , wherein determining the respective recommendation scores includes identifying, for a particular pair of brands, a common set of non-brand terms from the non-brand terms associated with each of the particular pair of brands. 
     
     
         4 . The method of  claim 2 , further comprising, for a particular brand of the plurality of brands, generating an association score for each non-brand term associated with the particular brand, wherein a given association score indicates a strength of association between the particular brand and a respective one of the associated non-brand terms. 
     
     
         5 . The method of  claim 4 , further comprising generating, for the particular brand, a vector representation using the association scores. 
     
     
         6 . The method of  claim 5 , wherein determining the respective recommendation score for the particular brand and a different brand includes calculating the respective recommendation score using the vector representation of the particular brand and a vector representation of the different brand. 
     
     
         7 . The method of  claim 4 , further comprising adding the particular brand to one or more brand categories using the association scores for the non-brand term associated with the particular brand. 
     
     
         8 . The method of  claim 2 , further comprising:
 identifying a group of queries that include a particular brand;   tokenizing the group of queries into respective sets of terms based on statistically significant phrases included in the group of queries; and   adding the particular brand to one or more brand categories based on the sets of terms.   
     
     
         9 . The method of  claim 2 , further comprising:
 identifying a set of queries issued by a particular user during a given session on a particular computing device; and   identifying a connection between two brands included in the set of queries.   
     
     
         10 . The method of  claim 9 , further comprising determining a strength of connection between the two brands based on identified connections between two brands across a plurality of user sessions. 
     
     
         11 . A non-transitory machine-readable medium having stored thereon machine-readable instructions executable to cause a machine to perform operations comprising:
 identifying a plurality of brands included in one or more of a plurality of user queries that are stored in a database;   identifying one or more non-brand terms associated with respective ones of the plurality of brands;   for pairs of brands of the plurality of brands, determining respective recommendation scores that are based on the associated non-brand terms for corresponding pairs of brands, wherein a given recommendation score is indicative of a strength of relationship between a respective pair of brands;   identifying a preferred brand for a user that is currently active on a computing device; and   selecting, based on the preferred brand and respective recommendation scores, one or more different brands as a recommendation to the user.   
     
     
         12 . The non-transitory machine-readable medium of  claim 11 , wherein the operations further comprise:
 for a particular brand of the plurality of brands, generating an association score for each non-brand term associated with the particular brand, wherein a given association score indicates a strength of association between the particular brand and a respective one of the associated non-brand terms; and   generating, for the particular brand, a vector representation using the association scores.   
     
     
         13 . The non-transitory machine-readable medium of  claim 12 , wherein determining the respective recommendation score for the particular brand and a different brand includes calculating the respective recommendation score using the vector representation of the particular brand and a vector representation of the different brand. 
     
     
         14 . The non-transitory machine-readable medium of  claim 12 , wherein the operations further comprise adding the particular brand to one or more brand categories using the association scores for the non-brand term associated with the particular brand. 
     
     
         15 . The non-transitory machine-readable medium of  claim 11 , wherein the operations further comprise:
 identifying a group of queries including a particular brand;   tokenizing the group of queries into respective sets of terms based on statistically significant phrases included in the group of queries; and   adding the particular brand to one or more brand categories based on the sets of terms.   
     
     
         16 . The non-transitory machine-readable medium of  claim 11 , wherein the operations further comprise:
 identifying a set of queries issued by a particular user during a given session on a particular computing device;   identifying a connection between two brands included in the set of queries; and   determining a strength of connection between the two brands based on identified connections between two brands across a plurality of user sessions.   
     
     
         17 . A system comprising:
 a database having stored therein a plurality of user queries;   a non-transitory memory; and   one or more hardware processors coupled to the non-transitory memory and configured to read instructions from the non-transitory memory to cause the system to perform operations comprising:
 identifying, from the database, a plurality of brands included in one or more of the plurality of user queries; 
 for respective ones of the plurality of brands, associating one or more non-brand terms included in the plurality of user queries; 
 for pairs of brands of the plurality of brands, determining respective recommendation scores based on the associated non-brand terms, wherein a given recommendation score is indicative of a strength of relationship between a respective pair of brands; 
 determining a preferred brand for a user that is currently active on a computing device that is coupled, via a network, to the system; and 
 based on the preferred brand and the respective recommendation scores, identifying one or more different brands for use as a recommendation to the user. 
   
     
     
         18 . The system of  claim 17 , wherein determining the respective recommendation scores includes identifying, for a particular pair of brands, a common set of non-brand terms from the non-brand terms associated with each of the particular pair of brands. 
     
     
         19 . The system of  claim 17 , wherein the operations further comprise generating, for a particular brand of the plurality of brands, an association score for each non-brand term associated with the particular brand, wherein a given association score indicates a strength of association between the particular brand and a respective one of the associated non-brand terms. 
     
     
         20 . The system of  claim 17 , wherein the operations further comprise:
 for a group of queries that include a particular brand, tokenizing ones of the group into respective sets of terms based on statistically significant phrases included in the group of queries; and   adding the particular brand to one or more brand categories based on the sets of terms.   
     
     
         21 . The system of  claim 17 , wherein the operations further comprise identifying, for a set of queries issued by a particular user during a given session on a particular computing device, a connection between two brands included in the set of queries.