Systems and methods for providing culturally-relevant search results to users
View Patent ↗Search results may be provided to a user. A search query may be received from the user. A query feature vector may be formed for the search query. The query feature vector may be compared with news feature vectors associated with documents related to current events. An augmented query feature vector may be formed based on results of the comparison of the query feature vector with the news feature vectors. The augmented query feature vector may be compared with feature vectors related to target documents. Search results that include target documents may be identified based on results of the comparison of the augmented query feature vector with the feature vectors related to the target documents. The user may be made able to perceive at least some of the identified search results.
1. A computer-implemented method, the method comprising the following operations performed by at least one processor:
forming a first feature vector for-a search query received from a user having a plurality of search terms and being associated with a particular date, the first feature vector comprising a first set of numerical values corresponding to the search terms and including user interest data;
obtaining a set of second feature vectors for a plurality of documents having publication dates that are temporally proximate to the particular date, the set of second feature vectors comprising second sets of numerical values associated with terms in corresponding ones of the documents;
computing, based on the second sets of numerical values, a centroid feature vector representative of at least a portion of the second feature vectors;
generating an augmented feature vector based on a comparison of the first feature vector and the centroid feature vector; and
identifying at least one target document that corresponds to the search query based on the augmented feature vector.
2. The method of claim 1 , further comprising receiving the search query from a device associated with the user, the search query being received from the user device on the particular date.
3. The method of claim 1 , wherein the plurality of documents comprise at least one document associated with a current event, the at least one current event document having a publication date that falls within a threshold time period of a date associated with the search query.
4. The method of claim 1 , wherein the computing comprises:
identifying one or more of the documents that are associated with at least one of the search terms;
determining a subset of the second feature vectors that correspond to the identified documents; and
generating the centroid feature vector based on the second sets of numerical value associated with the subset of the second feature vectors.
5. The method of claim 1 , wherein the computing comprises:
calculating, using the first and second numerical values, metrics of similarity between the first feature vector and the second feature vectors;
identifying a subset of the second feature vectors based on the calculated similarity metrics; and
generating the centroid feature vector based on the second sets of numerical value associated with the subset of the second feature vectors.
6. The method of claim 1 , wherein the generating comprises:
obtaining one or more of the documents associated with at least one of the search terms;
identifying a term within the obtained documents that is absent from the search terms; and
generating, for the augmented query feature vector, a set of numerical values corresponding to the search terms and the identified term.
7. The method of claim 5 , wherein the calculated similarity metrics comprise at least one of a distance or an angle between the first feature vector and a corresponding one of the second feature vectors in n-dimensional space.
8. The method of claim 1 , wherein the identifying comprises:
obtaining third feature vectors associated with a plurality of candidate target documents, the third feature vectors comprising third sets of numerical values associated with terms in corresponding ones of the candidate target documents;
computing metrics of similarity between the augmented query feature vector and the third feature vectors, the similarity metrics comprising at least one of distances or angles between the augmented query feature vector and the third feature vectors; and
selecting the at least one target document from the candidate target documents based on the computed similarity metrics.
9. The method of claim 1 , further comprising generating one or more electronic instructions to transmit information identifying the at least one target document to a device of a user.
10. The method of claim 1 , further comprising:
obtaining, based on the augmented feature vector, a plurality of target documents that correspond to the search query;
identifying a first portion of the target documents of relevance to a current event; and
generating one or more electronic instructions to transmit information identifying the current event and the first portion of the target documents to a device of a user.
11. An apparatus, comprising:
a storage device that stores a set of instructions; and
at least one processor coupled to the storage device, the at least one processor being operative with the set of instructions in order to:
form a first feature vector for a search query received from a user having a plurality of search terms and being associated with a particular date, the first feature vector comprising a first set of numerical values corresponding to the search terms and including user interest data;
obtain a set of second feature vectors for a plurality of documents having publication dates that are temporally proximate to the particular date, the set of second feature vectors comprising second sets of numerical values associated with terms in corresponding ones of the documents;
compute, based on the second sets of numerical values, a centroid feature vector representative of at least a portion the second feature vectors;
generate an augmented feature vector based on a comparison of the first feature vector and the centroid feature vector; and
identify at least one target document that corresponds to the search query based on the augmented feature vector.
12. The apparatus of claim 11 , wherein the at least one processor is further configured to receive the search query from a device associated with the user, the search query being received from the user device on the particular date.
13. The apparatus of claim 11 , wherein the plurality of documents comprise at least one document associated with a current event, the at least one current event document having a publication date that falls within a threshold time period of a date associated with the search query.
14. The apparatus of claim 11 , wherein the at least one processor is further configured to:
identify one or more of the documents that are associated with at least one of the search terms;
determine a subset of the second feature vectors that correspond to the identified documents; and
generate the centroid feature vector based on the second sets of numerical value associated with the subset of the second feature vectors.
15. The apparatus of claim 11 , wherein the at least one processor is further configured to:
calculate, using the first and second numerical values, metrics of similarity between the first feature vector and the second feature vectors;
identify a subset of the second feature vectors based on the calculated similarity metrics; and
generate the centroid feature vector based on the second sets of numerical value associated with the subset of the second feature vectors.
16. The apparatus of claim 11 , wherein the at least one processor is further configured to:
obtain one or more of the documents associated with at least one of the search terms;
identify a term within the obtained documents that is absent from the search terms; and
generate, for the augmented query feature vector, a set of numerical values corresponding to the search terms and the identified term.
17. The apparatus of claim 11 , wherein the at least one processor is further configured to:
obtain third feature vectors associated with a plurality of candidate target documents, the third feature vectors comprising third sets of numerical values associated with terms in corresponding ones of the candidate target documents;
compute metrics of similarity between the augmented query feature vector and the third feature vectors, the similarity metrics comprising at least one of distances or angles between the augmented query feature vector and the third feature vectors; and
select the at least one target document from the candidate target documents based on the computed similarity metrics.
18. The apparatus of claim 11 , wherein the at least one processor is further configured to generate one or more electronic instructions to transmit information identifying the at least one target document to a device of a user.
19. The apparatus of claim 11 , wherein the at least one processor is further configured to:
obtain, based on the augmented feature vector, a plurality of target documents that correspond to the search query;
identify a first portion of the target documents of relevance to a current event; and
generate one or more electronic instructions to transmit information identifying the current event and the first portion of the target documents to a device of a user.
20. A tangible, non-transitory computer-readable medium that stores a set of instructions that, when executed by at least one processor, cause the at least one processor to perform a method comprising:
forming a first feature vector for-a search query received from a user having a plurality of search terms and being associated with a particular date, the first feature vector comprising a first set of numerical values corresponding to the search terms and including user interest data;
obtaining a set of second feature vectors for a plurality of documents having publication dates that are temporally proximate to the particular date, the set of second feature vectors comprising second sets of numerical values associated with terms in corresponding ones of the documents;
computing, based on the second sets of numerical values, a centroid feature vector representative of at least a portion of the second feature vectors;
generating an augmented feature vector based on a comparison of the first feature vector and the centroid feature vector; and
identifying at least one target document that corresponds to the search query based on the augmented feature vector.
21. The method of claim 1 , wherein at least one of the search terms includes the particular date.