Method and apparatus for aligning multiple taxonomies

• US 8,732,197 B2
• Filed: 02/04/2008
• Issued: 05/20/2014
• Est. Priority Date: 02/02/2007
• Status: Active Grant

• Current Assignee: Callisto Publishing LLC
• Original Assignee: Musgrove Technology Enterprises, LLC
• Inventors: Musgrove, Timothy A.
• Primary Examiner(s): Boccio, Vincent F
• Application Number:
  US12/025,200
• Publication Number:
  US 20090037457A1
• Time in Patent Office: 2,297 Days
• Field of Search: 707/999.102, 707/705, 707/749, 707/755, 707/756, 707/776, 707/777
• US Class Current: 707/776
• CPC Class Codes: G06F 16/358 Browsing; Visualisation therefor
• CPC Class Codes: G06F 16/36 Creation of semantic tools, e.g. ontology or thesauri
• CPC Class Codes: G06F 16/367 Ontology

Abstract

A document taxonomy alignment system and method, relying on document glosses and utilizing a soft ontology expansion. An all-new hierarchical leaf node can be created expressly for the purpose of better aligning the plurality of document taxonomies in question. A small but valuable subset of the nodes created by soft ontology expansion turn out to capture some otherwise unmappable taxonomy nodes, and thereby have the effect of classifying the documents better than would any pre-existing node in any one of those taxonomies.

Claims

1. A method comprising:

mapping a first set of concept nodes in a first taxonomy to a second set of concept nodes in a second taxonomy by aligning nodes with equivalent concepts in the first taxonomy and second taxonomy to generate a master taxonomy having a plurality of mapped concept nodes, the first set of concept nodes for organizing a first plurality of documents and the second set of concept nodes for organizing a second plurality of documents, each of the first and second plurality of documents associated with a document gloss, each of the mapped concept nodes containing documents from a concept node from the first taxonomy and a concept node from the second taxonomy that are determined to contain equivalent categories of documents;

after mapping the first set of concept nodes to the second set of concept nodes,

finding an expanded concept that is instantiated disproportionately in the document glosses of an unmapped node of the first taxonomy;

determining if the expanded concept is instantiated in documents not classified at any leaf node in the second taxonomy; and

creating a new node with the expanded concept in the master taxonomy;

placing documents from the unmapped node and documents associated with the expanded concept from the second taxonomy not classified at any leaf node under the new node if the expanded concept is instantiated in documents not classified at any leaf node in the second taxonomy.

2. The method of claim 1 further comprising reviewing the first taxonomy and the second taxonomy.

3. The method of claim 1 wherein the nodes are categories of the first and second plurality of documents.

4. The method of claim 1 wherein the first and second plurality of documents are websites.

5. The method of claim 1 wherein document glosses are summaries of the first and second plurality of documents in the taxonomies and identify the differentiation of the document from other documents in the taxonomies.

6. The method of claim 1, wherein mapping the first set of concept nodes in the first taxonomy to the second set of concepts in the second taxonomy by aligning nodes with equivalent concepts in the first taxonomy and second taxonomy to generate the master taxonomy having the plurality of concept nodes comprises:

combining all documents of all nodes in each of the first and second taxonomies in a single set;

performing a semantic analysis of the documents and the document glosses;

if two words are frequently paired in the document glosses and in the documents, then the documents are related, wherein the word pairs are the concepts;

checking the node document glosses for the presence of the concepts; and

employing a percentage matching algorithm to connect node documents in one taxonomy to leaf nodes of a different taxonomy.

7. A computer readable storage medium having computer executable instructions recorded thereon which cause a computer system to carry out a method when executed, the method comprising:

mapping a first set of concept nodes in a first taxonomy to a second set of concepts in a second taxonomy by aligning nodes with equivalent concepts in the first taxonomy and second taxonomy to generate a master taxonomy having a plurality of mapped concept nodes, the first set of concept nodes for organizing a first plurality of documents and the second set of concept nodes for organizing a second plurality of documents, each of the first and second plurality of documents associated with a document gloss, each of the mapped concept nodes containing documents from a concept node from the first taxonomy and a concept node from the second taxonomy that are determined to contain equivalent categories for documents;

after mapping the first set of concept nodes to the second set of concept nodes,

finding an expanded concept that is instantiated disproportionately in the document glosses of an unmapped node of the first taxonomy;

determining if the expanded concept is instantiated in documents not classified at any leaf node in the second taxonomy;

creating a new node with the expanded concept in the master taxonomy; and

placing the documents from the unmapped node and documents associated with the expanded concept under the new node if the expanded concept from the second taxonomy not classified at any leaf node is instantiated in documents not classified at a leaf node in the second taxonomy.

8. The non-transitory computer readable storage medium of claim 7 further comprising reviewing the first taxonomy and the second taxonomy.

9. The non-transitory computer readable storage medium of claim 7 wherein the nodes are categories of the first and second plurality of documents.

10. The non-transitory computer readable storage medium of claim 7 wherein the first and second plurality of documents are websites.

11. The non-transitory computer readable storage medium of claim 7 wherein document glosses are summaries of the first and second plurality of documents in the taxonomies and identify the differentiation of the document from other documents in the taxonomies.

12. The non-transitory computer readable storage medium of claim 7, wherein mapping the first set of concept nodes in the first taxonomy to the second set of concepts in the second taxonomy by aligning nodes with equivalent concepts in the first taxonomy and second taxonomy to generate the master taxonomy having the plurality of concept nodes comprises:

combining all documents of all nodes in the first taxonomy and the second taxonomy in a single set;

performing a semantic analysis of the documents and the document glosses;

if two words are frequently paired in the document glosses and in the documents, then the documents are related, wherein the word pairs are the concepts; and

checking the node document glosses for the presence of the concepts; and

employing a percentage matching algorithm to connect node documents in one taxonomy to leaf nodes of a different taxonomy.

13. A computer system for mapping a taxonomy to at least one other taxonomy, wherein the computer system comprises a non-transitory computer readable storage medium having computer executable instructions recorded thereon which cause a computer system to carry out a method, the taxonomies including concepts for organizing information, the computer system comprising:

means for mapping a first set of concept nodes in a first taxonomy to a second set of concepts in a second taxonomy by aligning nodes with equivalent concepts in the first taxonomy and second taxonomy to generate a master taxonomy having a plurality of mapped concept nodes, the first set of concept nodes for organizing a first plurality of documents and the second set of concept nodes for organizing a second plurality of documents, each of the first and second plurality of documents associated with a document gloss, each of the mapped concept nodes containing documents from a concept node from the first taxonomy and a concept node from the second taxonomy that are determined to contain equivalent categories of documents;

after mapping the first set of concept nodes to the second set of concept nodes,

a concept finding module to find an expanded concept that is instantiated disproportionately in the document glosses of an unmapped node of the first taxonomy;

a test module to determine if the expanded concept is instantiated in documents not classified at a leaf node in the second taxonomy; and

a node creation module to create a new node with the expanded concept in the master taxonomy and place documents from the unmapped node and documents associated with the expanded concept under the new node if the expanded concept from the second taxonomy not classified at any leaf node is instantiated in documents not classified at a leaf node in the second taxonomy.

14. The system of claim 13 further comprising means for reviewing the first taxonomy and the second taxonomy.

15. The system of claim 13 wherein the nodes are categories of the first and second plurality of documents.

16. The system of claim 13 wherein the first and second plurality of documents are websites.

17. The system of claim 13 wherein document glosses are summaries of the first and second plurality of documents in the taxonomies and identify the differentiation of the document from other documents in the taxonomies.

18. The system of claim 13, wherein the means for mapping the first set of concept nodes in the first taxonomy to the second set of concepts in the second taxonomy by aligning nodes with equivalent concepts in the first taxonomy and second taxonomy to generate the master taxonomy having the plurality of concept nodes comprises means for:

combining all documents of all nodes in the first taxonomy and the second taxonomy in a single set;

performing a semantic analysis of the documents and the document glosses;

if two words are frequently paired in the document glosses and in the documents, then the documents are related, wherein the word pairs are the concepts; and

checking the node document glosses for the presence of the concepts; and

employing a percentage matching algorithm to connect node documents in one taxonomy to leaf nodes of a different taxonomy.