Selecting high quality text within identified reviews for display in review snippets
A method and system of selecting content from reviews for display are described. A review is identified and partitioned. A subset of the partitions is selected based on predefined criteria. A snippet is generated from the selected partitions. A response that includes the snippet is generated.
1. A computer implemented method of generating a review sample of a single review in a corpus of reviews, comprising:
at a server having one or more processors and memory storing one or more programs for execution by the one or more processors:
identifying the single review;
partitioning the review into partitions;
selecting a plurality of the partitions of the single review based on predefined criteria, wherein selecting the plurality of the partitions comprises: identifying two or more combinations of the partitions, determining a quality score for each respective combination, and selecting the respective combination having a highest quality score;
generating a snippet solely for the single review including content from each of the selected plurality of the partitions, wherein the review sample of the single review comprises the generated snippet; and
generating a response including the snippet for the single review.
2. The method of claim 1 , wherein selecting comprises:
determining a quality score for each of the partitions; and
selecting a plurality of the partitions of the single review based on the quality scores.
3. The method of claim 2 , wherein determining comprises determining a quality score for each of the partitions based on a length of the respective partition and at least one of the group consisting of: inverse document frequency values associated with one or more words in the respective partition, and a position of the respective partition within the review.
4. The method of claim 1 , wherein selecting comprises:
determining a quality score for each of the partitions; and
selecting a plurality of the partitions based on a combined quality score that is based on the quality scores.
5. The method of claim 1 , including limiting a length of the selected plurality of the partitions so that the length of the selected plurality of the partitions exceeds a predefined maximum snippet length by less than a full sentence.
6. A computer implemented method of processing reviews, comprising:
at a server having one or more processors and memory storing one or more programs for execution by the one or more processors:
identifying a single review;
partitioning the review into one or more partitions;
selecting a plurality of the partitions based on predefined criteria;
generating a snippet solely for the single review including content from the selected plurality of the partitions; and
generating a response including the snippet for the single review,
wherein selecting a plurality of the partitions comprises identifying one or more combinations of the partitions, each combination including a respective subset of the partitions that are consecutive within the review and satisfying predefined length criteria.
7. The method of claim 6 , wherein selecting a plurality of the partitions comprises:
determining a quality score for each partition in a respective combination; and
selecting the respective combination having a highest respective combination score, the respective combination score based on respective quality scores of the partitions in the respective combination.
8. A system for generating a review sample of a single review in a corpus of reviews, comprising:
one or more processors; and
memory storing one or more programs to be executed by the one or more processors, the one or more modules including instructions:
to identify the single review;
to partition the review into partitions;
to select a plurality of the partitions of the single review based on predefined criteria;
to generate a snippet solely for the single review including content from each of the selected plurality of the partitions, wherein the review sample of the single review comprises the generated snippet; and
to generate a response including the snippet for the single review,
wherein the instructions to select the plurality of the partitions include instructions to: identify two or more combinations of the partitions, determine a quality score for each respective combination, and select the respective combination having a highest quality score.
9. The system of claim 8 , wherein the one more modules include instructions:
to determine a quality score for each of the partitions; and
to select a plurality of the partitions of the single review based on the quality scores.
10. The system of claim 9 , wherein the one or more modules include instructions to determine a quality score for each of the partitions based on a length of the respective partition and at least one of the group consisting of: inverse document frequency values associated with one or more words in the respective partition, and a position of the respective partition within the review.
11. The system of claim 8 , wherein the one more modules include instructions:
to determine a quality score for each of the partitions;
to select a plurality of the partitions based on a combined quality score that is based on the quality scores.
12. A system for processing reviews, comprising:
one or more processors; and
memory storing one or more programs to be executed by the one or more processors, the one or more modules including instructions:
to identify a single review;
to partition the review into one or more partitions;
to select a plurality of the partitions based on predefined criteria;
to generate a snippet solely for the single review including content from the selected plurality of the partitions; and
to generate a response including the snippet for the single review,
wherein the one or more modules include instructions to identify one or more combinations of the partitions, each combination including a respective plurality of the partitions that are consecutive within the review and satisfying predefined length criteria.
13. The system of claim 12 , wherein the one or more modules include instructions:
to determine a quality score for each partition in a respective combination; and
to select the respective combination having a highest respective combination score, the respective combination score based on respective quality scores of the partitions in the respective combination.
14. A non-transitory computer readable storage medium storing one or more programs for execution by one or more processors at a server computer, the one or more programs comprising instructions for:
identifying a single review;
partitioning the review into partitions;
selecting a plurality of the partitions of the single review based on predefined criteria, wherein selecting the plurality of the partitions comprises: identifying two or more combinations of the partitions, determining a quality score for each respective combination, and selecting the respective combination having a highest quality score;
generating a snippet solely for the single review including content from each of the selected plurality of the partitions; and
generating a response including the snippet for the single review.
15. The computer readable storage medium of claim 14 , wherein the instructions for selecting comprise instructions for:
determining a quality score for each of the partitions; and
selecting a plurality of the partitions of the single review based on the quality scores.
16. The computer readable storage medium of claim 15 , wherein the instructions for determining comprise instructions for determining a quality score for each of the partitions based on a length of the respective partition and at least one of the group consisting of: inverse document frequency values associated with one or more words in the respective partition, and a position of the respective partition within the review.
17. The computer readable storage medium of claim 14 , wherein instructions for selecting comprise:
instructions for determining a quality score for each of the partitions; and
instructions for selecting a plurality of the partitions based on a combined quality score that is based on the quality scores.
18. A non-transitory computer readable storage medium storing one or more programs for execution by one or more processors at a server computer, the one or more programs comprising instructions for:
identifying a single review;
partitioning the review into one or more partitions;
selecting a plurality of the partitions based on predefined criteria;
generating a snippet solely for the single review including content from each of the selected plurality of the partitions; and
generating a response including the snippet for the single review,
wherein the instructions for selecting a plurality of the partitions comprise identifying one or more combinations of the partitions, each combination including a respective subset of the partitions that are consecutive within the review and satisfying predefined length criteria.
19. The computer readable storage medium of claim 12 , wherein the instructions for selecting a plurality of the partitions based on the quality scores comprise instructions for:
determining a quality score for each partition in a respective combination; and
selecting the respective combination having a highest respective combination score, the respective combination score based on respective quality scores of the partitions in the respective combination.