Computer-implemented system and method for generating a reference set via seed documents
View Patent ↗A computer-implemented system and method for generating a reference set via seed documents is provided. A collection of documents is obtained. One or more seed documents are identified. The seed documents are compared with the document collection and those documents that are similar to the seed documents are identified as reference set candidates. A size threshold is applied to the reference set candidates, which are grouped as the reference set when the size threshold is satisfied.
1. A computer-implemented method for generating a reference set via seed documents, comprising:
obtaining a collection of documents;
identifying one or more seed documents related to an issue for which a reference set of documents is needed;
selecting the seed documents from at least one of a current document set and a previously defined document set;
comparing the seed documents to the document collection and identifying those documents similar to the seed documents as reference set candidates;
applying a size threshold to the reference set candidates; and
grouping the reference set candidates as the reference set when the size threshold is satisfied, wherein the reference set is reduced by identifying the reference set candidates that are closely related and common reference set candidates and by removing the closely related reference set candidates and the common reference set candidates.
2. A method according to claim 1 , further comprising:
refining the reference set by reducing a number of the reference set candidates included in the reference set when the size threshold is not satisfied.
3. A method according to claim 1 , further comprising:
determining the seed documents using a keyword search.
4. A method according to claim 1 , further comprising:
identifying documents from the collection that are dissimilar to the seed documents; and
combining the similar documents and the dissimilar documents as the reference set candidates.
5. A method according to claim 1 , further comprising:
determining the seed documents, comprising:
identifying a topic of the document collection; and
identifying documents that are important to the topic from one of the document collection and the a further collection.
6. A method according to claim 1 , further comprising:
grouping the collection of documents into assignments.
7. A method according to claim 1 , further comprising at least one of:
assigning a classification code to one or more of the reference set candidates in the reference set; and
receiving a classification code for one or more of the reference set candidates in the reference set.
8. A method according to claim 1 , further comprising:
determining features of the documents in the collection;
obtaining seed features and applying the seed features to the determined features;
identifying those features that are similar to the seed features as reference set candidate features; and
adding the documents associated with the reference set candidate features to the reference set.
9. A computer-implemented system for generating a reference set via seed documents, comprising:
a document module to obtain a collection of documents;
a seed identification module to identify one or more seed documents related to an issue for which a reference set of documents is needed;
a seed selection module to select the seed documents from at least one of a current document set and a previously defined document set;
a comparison module to compare the seed documents to the document collection and to identify those documents similar to the seed documents as reference set candidates;
a size module to apply a size threshold to the reference set candidates;
a reference set module to group the reference set candidates as a reference set when the size threshold is satisfied, wherein the reference set is reduced by identifying the reference set candidates that are closely related and common reference set candidates and by removing the closely related reference set candidates and the common reference set candidates; and
a processor to execute the modules.
10. A system according to claim 9 , further comprising:
a refinement module to refine the reference set by reducing a number of the reference set candidates included in the reference set when the size threshold is not satisfied.
11. A system according to claim 9 , further comprising:
a seed determination module to determine the seed documents using a keyword search.
12. A system according to claim 9 , further comprising:
a dissimilarity module to identify documents from the collection that are dissimilar to the seed documents; and
the reference set module to combine the similar documents and the dissimilar documents as the reference set candidates.
13. A system according to claim 9 , further comprising:
a seed determination module to determine the seed documents, comprising:
a topic module to identify a topic of the document collection; and
a document importance module to identify documents from at least one of the document collection and a further collection that are important to the topic.
14. A system according to claim 9 , further comprising:
an assignment module to group the collection of documents into assignments.
15. A system according to claim 9 , further comprising at least one of:
an automatic classification module to assign a classification code to one or more of the reference set candidates in the reference set; and
a classification receipt module to receive a classification code for one or more of the reference set candidates in the reference set.
16. A system according to claim 9 , further comprising:
a feature identification module to determine features of the documents in the collection;
a seed feature module to obtain seed features and to apply the seed features to the determined features;
a feature similarity module to identify those features that are similar to the seed features as reference set candidate features; and
the reference set module to include the documents associated with the reference set candidate features with the reference set.
17. A computer-implemented method for generating a reference set via seed documents, comprising:
obtaining a collection of documents;
identifying one or more seed documents related to an issue for which a reference set of documents is needed;
comparing the seed documents to the document collection and identifying those documents similar to the seed documents as reference set candidates;
applying a size threshold to the reference set candidates;
grouping the reference set candidates as the reference set when the size threshold is satisfied;
determining features of the documents in the collection;
obtaining seed features and applying the seed features to the determined features;
identifying those features that are similar to the seed features as reference set candidate features; and
adding the documents associated with the reference set candidate features to the reference set.
18. A method according to claim 17 , further comprising:
refining the reference set by reducing a number of the reference set candidates included in the reference set when the size threshold is not satisfied.
19. A method according to claim 17 , further comprising:
determining the seed documents using a keyword search.
20. A method according to claim 17 , further comprising:
determining the seed documents, comprising:
identifying a topic of the document collection; and
identifying documents that are important to the topic from one of the document collection and the a further collection.