Systems and methods for employing an orthogonal corpus for document indexing

• US 7,275,061 B1
• Filed: 04/13/2000
• Issued: 09/25/2007
• Est. Priority Date: 04/13/2000
• Status: Expired due to Fees

• Current Assignee: Linkapedia, Inc.
• Original Assignee: Indraweb.Com, Inc.
• Inventors: Burch, George, Kon, Henry
• Primary Examiner(s): Alam; Hosain
• Assistant Examiner(s): Woo; Isaac
• Application Number:
  US09/548,796
• Time in Patent Office: 2,721 Days
• Field of Search: 707 1- 10, 707100-1041, 707200-206, 715530-541
• US Class Current: 1/1
• CPC Class Codes: G06F 16/353 into predefined classes
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06Q 30/0601 Electronic shopping [e-shopping]
• CPC Class Codes: Y10S 707/99942 Manipulating data structure, e.g. compression, compaction, compilation
• CPC Class Codes: Y10S 707/99943 Generating database or data structure, e.g. via user interface
• CPC Class Codes: Y10S 707/99944 Object-oriented database structure
• CPC Class Codes: Y10S 707/99945 Object-oriented database structure processing

Abstract

The invention provides for indexing and cataloging of content on the Internet, as well as from other stores of information, may be performed by applying a process that employs an orthogonal corpus, or corpora, of information, such as an Encyclopedia. To this end, the processes described herein identify the topics discussed within the corpus. The process also identifies within the corpus a set of keywords that are relevant to the topics presented in the corpus. The keywords associated with a topic may be employed to identify documents stored in another database that are related to the topic. A graphical representation of the index of topics found in the corpus may then be generated, with individual topics operating as links to these related documents. Thus, a user interested in reviewing content in the corpus related to a certain topic, may also activate a link in the graphical representation of the index to access other documents that have been identified as related to the topic of interest to the user.

Claims

1. A method for topical indexing a document collection that is initially unconnected with a body of textual reference material, comprising:

processing the body of textual reference material into a plurality of text portions, each text portion being associated with a single topic from a plurality of topics,

processing said plurality of text portions to derive keywords for each topic,

assigning a weight to each keyword in a text portion;

associating a keyword with a corresponding text portion if the weight of said keyword in said corresponding text portion is equal to or greater than a weight of said keyword in the text portions other than the corresponding text portion, or is equal to or greater than a predetermined threshold value;

forming first keyword-weight pairs of said associated keywords;

applying the associated keywords to at least one document from said initially unconnected document collection and forming second keyword-weight pairs associated with said at least one document;

forming a numeric score between the first and second keyword-weight pairs; and

based on said score, associating said at least one document from said initially unconnected document collection and said single topic from said plurality of topics.

2. A method according to claim 1, wherein the body of textual reference material is selected from a group consisting of an encyclopedia, a dictionary, a text book, a novel, a newspaper, a web site, and a glossary.

3. A method according to claim 1, wherein the plurality of topics includes a hierarchical organization of substantially orthogonal topics.

4. A method according to claim 1, wherein processing the textual reference material includes identifying a table of contents and chapter headings for the textual reference material.

5. A method according to claim 1, wherein processing the textual reference material includes identifying a keyword index for the body of textual reference material.

6. A method according to claim 1, wherein processing the textual reference material includes identifying definition entries in a dictionary or glossary.

7. A method according to claim 1, wherein processing the associated text portion includes generating a word map representative of a statistical analysis of words contained in the associated text portion.

8. A method according to claim 7, wherein generating the word map includes performing a word count process for determining word frequency of a word within the associated text portion, and employing the word frequency for determining relevance of the word to the associated text portion.

9. A method according to claim 1, wherein processing the associated text portion includes processing across a plurality of associated text portions to identify a set of keywords having an associated measure of intra-topic orthogonality.

10. A method according to claim 1, wherein processing the associated portion includes identifying a set of synonyms for extending the keywords.

11. A method according to claim 1, wherein processing the associated text portion includes extracting a subset of the keywords representing search keys having a predetermined measure of correlation to the single topic and applying the search keys to the at least one document before scoring the at least one document.

12. A method according to claim 11, wherein applying the search keys to the at least one document includes applying the search keys to an Internet search engine to discover relevant documents related to the search keys.

13. A method according to claim 11, wherein applying the search keys to the at least one document includes applying the search keys to a meta-search engine to discover relevant documents related to the search keys.

14. A method according to claim 1, wherein creating an association between the single topic and the at least one document includes capturing a location pointer associated with the at least one document.

15. A method according to claim 1, wherein creating an association between the single topic and the at least one document includes creating a data structure for the single topic for storing a location pointer associated with the at least one document.