Methods and computer-program products for organizing electronic documents

• US 10,127,229 B2
• Filed: 04/23/2014
• Issued: 11/13/2018
• Est. Priority Date: 04/23/2014
• Status: Active Grant

• Current Assignee: Elsevier BV (RELX PLC)
• Original Assignee: Elsevier BV (RELX PLC)
• Inventors: Schijvenaars, Bob, Doornenbal, Marius
• Primary Examiner(s): Le, Uyen T
• Application Number:
  US14/259,671
• Publication Number:
  US 20150310000A1
• Time in Patent Office: 1,665 Days
• CPC Class Codes: G06F 16/353 into predefined classes
• CPC Class Codes: G06F 16/355 Class or cluster creation or modification
• CPC Class Codes: G06F 16/382 using citations hypermedia G06F16/94
• CPC Class Codes: G06F 16/93 Document management systems
• CPC Class Codes: G06F 17/30011 Document retrieval systems
• CPC Class Codes: G06F 17/30707 into predefined classes
• CPC Class Codes: G06F 17/3071 including class or cluster creation or modification
• CPC Class Codes: G06F 17/30728 using citations hypermedia G06F17/30014

Abstract

Methods of organizing documents by reclassification and clustering are disclosed. In one embodiment, a method of clustering electronic documents of a document corpus includes comparing, by a computer, each individual electronic document in the document corpus with each other electronic document in the document corpus, thereby forming document pairs. The electronic documents of the document pairs are compared by calculating a similarity value with respect to the electronic documents of a document pair, associating the similarity value with both electronic documents of the document pair, and applying a clustering algorithm to the document corpus using the similarity values to create a plurality of hierarchical clusters. The similarity value is based on a plurality of attributes of the electronic documents in the document corpus. The plurality of attributes includes a citation attribute, a text-based attribute and one or more of an author-attribute, a publication-attribute, an institution-attribute, a downloads-attribute, and a clustering-results-attribute.

Claims

1. A computer-based method of reclassifying and clustering electronic documents of a document corpus to improve classification of the electronic documents so that correct documents are returned as a result of a computer-based search, the method comprising:

comparing, by a computer, each individual electronic document in the document corpus with each other electronic document in the document corpus, thereby forming document pairs, wherein the electronic documents of the document pairs are compared by;

calculating a similarity value with respect to the electronic documents of a document pair from a plurality of attributes of the electronic documents in the document corpus, the plurality of attributes comprising a citation attribute, a text-based attribute, and one or more of the following attributes;

an author attribute expressed as

2. The computer-based method of claim 1, wherein the citation attribute is based on citations between the electronic documents of the document pair and/or between the other electronic documents in the document corpus, and the text-based attribute is based on keywords extracted from the electronic documents of the document corpus.

3. The computer-based method of claim 1, wherein calculating the similarity value comprises generating a keyword vector for each electronic document in the document corpus from keywords extracted from the electronic documents of the document corpus.

4. The computer-based method of claim 3, further comprising extracting, from at least one thesaurus, one or more synonyms of one or more keywords extracted from the electronic documents of the document corpus, wherein the text-based attribute is further based on the synonyms.

5. The computer-based method of claim 1, wherein the citation attribute is based on one or more of direct citations, co-citations and bibliographic couplings.

6. The computer-based method of claim 5, wherein different weights are assigned to the one or more of the direct citations, co-citations and bibliographic couplings in calculating the similarity value.

7. The computer-based method of claim 1, wherein:

the citation attribute is based on direct citations, co-citations, and bibliographic couplings;

a weight associated with the direct citations is greater than a weight associated with the co-citations; and

the weight associated with the co-citations is greater than a weight associated with the bibliographic couplings.

8. The computer-based method of claim 1, further comprising extracting sample keywords from electronic documents within each cluster of the plurality of hierarchical clusters, and associating the sample keywords with each cluster.

9. A method of improving computer-based search results by organizing electronic documents of a document corpus, each one of the electronic documents comprising a predetermined classification code and having similarity values based on a similarity with respect to other electronic documents in the document corpus, the method comprising:

comparing each individual electronic document in the document corpus with each other electronic document in the document corpus, thereby forming document pairs, wherein the electronic documents of the document pairs are compared by;

calculating a similarity value with respect to the electronic documents of a document pair from a plurality of attributes of the electronic documents in the document corpus, the plurality of attributes comprising a citation attribute, a text-based attribute, and one or more of the following attributes;

an author attribute expressed as

10. The method of claim 9, wherein the citation attribute is based on direct citations, co-citations and bibliographic couplings.

11. The method of claim 9, wherein a weight associated with direct citations is greater than a weight associated with co-citation, and the weight associated with co-citation is greater than a weight associated with bibliographic coupling.

12. The method of claim 11, further comprising extracting, from at least one thesaurus, one or more synonyms of one or more keywords extracted from the electronic documents of the document corpus, wherein the text-based attribute is further based on the synonyms.

13. The method of claim 9, wherein the plurality of attributes further includes one or more of the following:

an author attribute, a publication attribute, an institution attribute, a downloads attribute, and a clustering results attribute.

14. The method of claim 9, wherein the classifier is a k-nearest neighbor classifier, and the electronic documents in the predetermined classification code represent a training set.

15. The method of claim 9, wherein an electronic document is reclassified to the evaluation classification code if the evaluation classification code is different from the predetermined classification code originally associated with the electronic document.

16. A computer-program product for clustering electronic documents of a document corpus to improve classification of the electronic documents so that correct documents are returned as a result of a computer-based search, the computer-program product comprising a non-transitory computer-readable medium storing executable instructions that, when executed by a computing device, cause the computing device to:

compare each individual electronic document in the document corpus with each other electronic document in the document corpus, thereby forming document pairs, wherein the electronic documents of the document pairs are compared by;

calculating a similarity value with respect to the electronic documents of a document pair from a plurality of attributes of the electronic documents in the document corpus, the plurality of attributes comprising a citation attribute, a text-based attribute and one or more of the following attributes;

an author attribute expressed as