Method and system for enhanced query term suggestion
Method, system, and programs for providing enhanced query term suggestions. Candidate query terms may be obtained based on a prefix of incomplete query terms received. The candidate query terms may be ranked, at least partially based on, their similarities with respect to query terms previously entered in the same search session as the incomplete query term. In some implementations, for determining such similarities, feature vectors and/or signatures may be stored in association with query terms. Similarity between a candidate query term and query terms in the same search session as the incomplete query term may be determined using the feature vectors and/or signatures associated therewith.
1. A method, implemented on a machine having at least one processor, storage, and a communication platform connected to a network, for providing query term suggestions, the method comprising:
receiving from a user, a partial query term in a search session;
obtaining, one or more candidate query terms based on the partial query term, and a plurality of query terms previously entered by the user, each query term of the plurality of query terms being included in the search session;
obtaining, a signature of the one or more candidate query terms, and each query term of the plurality of query terms, respectively;
computing an overall score for each of the one or more candidate query terms, the overall score being computed based on a degree of correlation between the signature of the candidate query term and each signature of the query term of the plurality of query terms; and
providing the one or more candidate query terms as query suggestions based on the overall scores of the candidate query terms.
2. The method of claim 1 , wherein obtaining the signature of the candidate query term, further comprises:
obtaining a feature vector including one or more features associated with the candidate query term; and
compressing the feature vector to generate the signature of the candidate query term.
3. The method of claim 2 , wherein the compressing step further comprises:
transforming a first representation of the feature vector in a first dimensional space to a second representation corresponding to the signature in a second dimensional space.
4. The method of claim 3 , wherein the first dimensional space is greater than the second dimensional space.
5. The method of claim 1 , wherein the overall score of a candidate query term is based on a weighted sum of correlations between the signature of the candidate term and each signature of the query term of the plurality of query terms.
6. The method of claim 1 , wherein the degree of correlation between the signature of the candidate query term and each signature of the query term of the plurality of query terms is computed based on a cosine-similarity function.
7. The method of claim 1 , further comprising:
ranking the one or more candidate query terms based on the computed overall score associated with each candidate query term.
8. The method of claim 1 , wherein the search session corresponds to a predetermined time-period with respect to a point of time the partial query term is received.
9. The method of claim 1 , wherein the signature is a locality-sensitive hash signature.
10. A system configured for providing query term suggestions, the system comprising:
at least one processor configured to:
receive from a user, a partial query term in a search session,
obtain, one or more candidate query terms based on the partial query term, and a plurality of query terms previously entered by the user, each query term of the plurality of query terms being included in the search session,
obtain, a signature of the one or more candidate query terms, and each query term of the plurality of query terms, respectively,
compute an overall score for each of the one or more candidate query terms, the overall score being computed based on a degree of correlation between the signature of the candidate query term and each signature of the query term of the plurality of query terms, and
provide the one or more candidate query terms as query suggestions based on the overall scores of the candidate query terms.
11. The system of claim 10 , wherein the at least one processor is further configured to:
obtain a feature vector including one or more features associated with the candidate query term; and
compress the feature vector to generate the signature of the candidate query term.
12. The system of claim 11 , wherein the wherein the at least one processor is further configured to:
transforming a first representation of the feature vector in a first dimensional space to a second representation corresponding to the signature in a second dimensional space, the first dimensional space being greater than the second dimensional space.
13. The system of claim 10 , wherein the overall score of a candidate query term is based on a weighted sum of correlations between the signature of the candidate term and each signature of the query term of the plurality of query terms.
14. The system of claim 10 , wherein the degree of correlation between the signature of the candidate query term and each signature of the query term of the plurality of query terms is computed based on a cosine-similarity function.
15. A non-transitory machine-readable medium having information recorded thereon for providing search term suggestions, wherein the information, when read by a machine causes the machine to perform the steps of:
receiving from a user, a partial query term in a search session;
obtaining, one or more candidate query terms based on the partial query term, and a plurality of query terms previously entered by the user, each query term of the plurality of query terms being included in the search session;
obtaining, a signature of the one or more candidate query terms, and each query term of the plurality of query terms, respectively;
computing an overall score for each of the one or more candidate query terms, the overall score being computed based on a degree of correlation between the signature of the candidate query term and each signature of the query term of the plurality of query terms; and
providing the one or more candidate query terms as query suggestions based on the overall scores of the candidate query terms.
16. The non-transitory machine-readable medium of claim 15 , further comprising:
obtaining a feature vector including one or more features associated with the candidate query term; and
compressing the feature vector to generate the signature of the candidate query term.
17. The non-transitory machine-readable medium of claim 16 , further comprising:
transforming a first representation of the feature vector in a first dimensional space to a second representation corresponding to the signature in a second dimensional space, the first dimensional space being greater than the second dimensional space.
18. The non-transitory machine-readable medium of claim 15 , wherein the overall score of a candidate query term is based on a weighted sum of correlations between the signature of the candidate term and each signature of the query term of the plurality of query terms.
19. The non-transitory machine-readable medium of claim 15 , wherein the degree of correlation between the signature of the candidate query term and each signature of the query term of the plurality of query terms is computed based on a cosine-similarity function.
20. The non-transitory machine-readable medium of claim 15 , wherein the search session corresponds to a predetermined time-period with respect to a point of time the partial query term is received.