Method and system for linking documents with multiple topics to related documents

• US 7,814,102 B2
• Filed: 12/07/2005
• Issued: 10/12/2010
• Est. Priority Date: 12/07/2005
• Status: Active Grant

• Current Assignee: RELX Inc. (RELX PLC)
• Original Assignee: LexisNexis Group Inc. (RELX PLC)
• Inventors: Miller, David James
• Primary Examiner(s): Rones; Charles
• Assistant Examiner(s): CAO, PHUONG THAO
• Application Number:
  US11/295,531
• Publication Number:
  US 20070130100A1
• Time in Patent Office: 1,770 Days
• Field of Search: 707/1, 707/999.001, 707/999.002, 707/999.003
• US Class Current: 707/736
• CPC Class Codes: G06F 16/382 using citations hypermedia G06F16/94
• CPC Class Codes: G06F 16/94 Hypermedia Hyperlinking G06F40/134

Abstract

A method and system for linking documents with multiple topics to related documents employs two distinct processes: a fabrication process to pre-establish candidate target document links within a case law or other document for discrete topics, for example, points of law and landmark cases, that represent related documents that are relevant to one or more headnotes within the case law document; and an on-line process to analyze the search request, the candidate links, and available landmark cases during a retrieval of a document to dynamically select the candidate target document links that are relevant to the point of law of interest to a searcher who retrieved the case law document via a search request.

Claims

1. A computer-implemented method for linking documents with multiple topics to related documents, implemented using a client/server network architecture, comprising:

in advance of a user search request, searching in a seed source to identify seed documents having at least one discrete seed topic but lacking target document links, using the client/server network architecture, each seed document belonging to one of a plurality of collections of seed documents, each collection having a target source map defining a list of target sources to search to generate target document links for seed documents belonging in the collection, wherein the at least one seed topic corresponds to a topical classification assigned from a taxonomy for a specific subject area, and wherein the at least one seed topic is pre-defined in advance of a user search request;

retrieving and formatting the identified seed documents, in advance of a user search request, using the client/server network architecture;

extracting the at least one seed topic for each retrieved seed document, in advance of a user search request, using the client/server network architecture;

for each retrieved seed document, retrieving the list of the target sources found in the target source map for the collection to which the seed document belongs, in advance of a user search request, using the client/server network architecture; and

for each seed topic within each retrieved seed document, processing each target source within the list of target sources defined by the target source map for the collection to which the retrieved seed document belongs, to pre-establish a target document link for the seed topic using a natural language search constructed from the seed topic, in advance of a user search request, using the client/server network architecture.

2. The method of claim 1, further comprising the steps of:

in response to retrieval of a seed document as the result of a search request from a user, selecting seed topics of the retrieved seed document that are relevant to the search request, based on terms of the search request occurring within the seed topics, whereby the selection of the seed topics is dynamic; and

displaying to the user those candidate target document links pre-established for the selected seed topics, whereby the displayed target document links are relevant to the retrieved seed document and to the search request resulting in retrieval of the retrieved seed document.

3. The method of claim 1, further comprising the step of preparing a source to be a target source by preparing documents in the source to be target documents in the target source, by assigning basic definitions to the target source, wherein the basic definitions include a definition of a portion of each of the target documents suitable as a description of the target document for purposes of displaying a link to a user, and a definition for generating a term vector for the target document, prior to the searching step.

4. The method of claim 3, wherein the step of preparing documents in a source includes the step of assigning basic definitions to the source.

5. The method of claim 1, wherein in the step of processing each target source, each target source within the list of target sources defined by the target source map for the retrieved seed document is processed for a document that qualifies for the current seed topic.

6. The method of claim 3, wherein each of the target documents in the target source has a citation, and wherein in the step of preparing documents in a source to be target documents, the basic definitions are implemented by a list of XML elements to retrieve for a display of a target document link in response to a user request and an XML element that specifies the citation for the target document.

7. The method of claim 6, wherein in the step of retrieving and formatting, the formatting includes extracting XML elements specific to seed topics of the identified seed documents, and wherein the step of processing each target source for each seed topic comprises the steps of:

constructing a special target document link search comprising a plurality of search terms, from the XML element for the seed topic;

running the special target document link search in a target source specified in the list of target sources;

returning target documents matching the special target document link search, from the specified target source;

performing a target document link retrieval request for each target document returned by the returning step;

in response to the target document link retrieval request, retrieving a requested target document;

loading the retrieved target document into an XML parser;

retrieving metadata for the retrieved target document based at least in part upon the target source;

storing the metadata in a data set;

generating core terms from the retrieved target document, based on the metadata;

locating and marking each of the search terms found within the retrieved target document;

comparing the found search terms within the retrieved target document with the core terms generated therefrom;

determining a score equal to the number of unique search terms that are also core terms;

generating an XML document containing all the data as defined within display of the target document link in response to the user request and the score;

comparing the scores returned for each retrieved target document;

retaining only those retrieved target documents meeting a pre-determined score threshold;

collecting all the retrieved target documents retained by the retaining step;

generating a new XML element containing the target document links; and

adding the new XML element back into the original seed document.

8. The method of claim 7, wherein the step of retrieving a requested target document in response to the target document link retrieval request, returns a citation for any document retrieved from the target source that is sufficient to generate a hypertext link to that document.

9. The method of claim 8, wherein the step of retrieving a requested target document in response to the target document link retrieval request, further returns suitable text to display as a label for the hypertext link.

10. The method of claim 7, wherein in the step of storing the metadata, the data set is indexed by source and view.

11. The method of claim 7, wherein in the step of generating core terms, the core terms represent the statistically most important words and phrases within the retrieved target document.

12. The method of claim 7, wherein the step of generating an XML document is carried out by applying an XSLT style sheet to select only the required XML elements within the retrieved target document and place them in new XML elements as required by a retrieving application.

13. The method of claim 7, wherein in the retaining step, only as many documents as are specified in a separate configurable value are retained.

14. The method of claim 12, wherein if tie scores occur, the retrieved target document ranked higher by an initial search is retained.

15. The method of claim 1, further comprising the steps of:

creating a list of landmark documents;

storing in a table the landmark documents, data representing the names of the landmark documents, the primary citations for the landmark documents, and the frequency with which each of the landmark documents is cited by other documents; and

creating a file containing the data in a format optimized for on-line retrieval.

16. The method of claim 1, wherein the step of processing each seed topic within each seed document to construct a target document link for each seed topic includes the steps of:

in response to a target document link search carried out in a target source specified in the list of target sources, returning matching target documents from the specified target source;

retaining only those matching target documents meeting pre-determined specifications;

generating a new XML element containing target document links for the retained target documents; and

adding the new XML element back into the original seed document.

17. The method of claim 16, further comprising the step of running multiple target document link searches for each extracted seed topic being processed.

18. The method of claim 7, further comprising the steps of:

in response to retrieval of a seed document as the result of a search request by a user, selecting the seed topics within the seed document that are relevant to the search request based on terms of the search request occurring within the seed topics;

following the step of selecting the seed topics, selecting target document links that have been fabricated for the selected seed topics;

for each seed topic selected as relevant to the search that caused the seed document to be retrieved, scoring the selected target document links fabricated into the seed document for the selected topics;

selecting a preset maximum number of selected target document links to be displayed as target document links, based on the type of publication of each target source;

for each target document link selected for display, constructing a hypertext link using the data previously stored in the XML element associated with the target document link in the XML document; and

organizing the hypertext links for display within the context of the seed document display.

19. The method of claim 18, wherein the step of selecting the seed topics comprises analyzing the search terms that occurred within the seed topics using traditional information retrieval techniques.

20. The method of claim 19, wherein the information retrieval techniques include at least one of term frequency, search term coverage, and seed topic length.

21. The method of claim 18, wherein in the scoring step, the score is generated based upon the number of unique search request terms within the target document that are also core terms of the selected target document link.

22. The method of claim 18, wherein in the step of selecting a preset maximum number of target document links, the selection is based upon the score assigned to the target document link and a business priority assigned to the target source.

23. The method of claim 22, wherein in the step of organizing the hypertext links for display, the hypertext links are organized by material type, and ordered within each material type by the assigned score and business priority.

24. The method of claim 18, wherein each hypertext link to the target document also contains a reference to the seed topic within the seed document for which the target document was selected.

25. The method of claim 24, wherein the reference to the seed topic is in the form of an intra-document link.