Search system and method

• US 7,668,825 B2
• Filed: 08/26/2005
• Issued: 02/23/2010
• Est. Priority Date: 08/26/2005
• Status: Active Grant

• Current Assignee: Seekr Technologies Incorporated
• Original Assignee: Convera Corporation
• Inventors: Vogel, Claude, Van Winkle, Jon Michael, Germundson, Eric, Powers, Joshua Michael, Robertson, Joel Wayne, Gardner, Jr., Paul
• Primary Examiner(s): LEROUX, ETIENNE PIERRE
• Application Number:
  US11/212,523
• Publication Number:
  US 20070050393A1
• Time in Patent Office: 1,642 Days
• Field of Search: 707/3, 707/4, 707/5, 707/102, 707/6
• US Class Current: 1/1
• CPC Class Codes: G06F 16/367 Ontology
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

A search system and method are provided that uses taxonomies, entities, facets, and ontologies to provide a user with a more comprehensive set of search results in response to a query. The search system has an indexing engine that performs one or more indexing steps that permit the search engine to return a comprehensive set of search results. For example, the indexing engine may index a document according to a set of synsets so that the search engine may use the synsets, during retrieval of results to a query, to return a more comprehensive set of search results.

Claims

1. A query processing system for processing a query having one or more query terms using formally represented knowledge against a corpus of documents, the system comprising:

a knowledge base containing a plurality of pieces of formally represented knowledge, wherein each piece of formally represented knowledge further comprises an item that has been edited or analyzed, the formally represented knowledge further comprising one or more synsets wherein each synset contains a group of terms that have a same meaning, one or more taxonomies wherein each taxonomy contains one or more synsets in a subject matter area that are organized from a synset having a general meaning to a synset having a specific meaning, one or more ontologies wherein each ontology contains one or more synsets associated with an area of interest and one or more facets wherein each facet is associated with a particular ontology and wherein a document is associated with the facet when the document contains the one or more synsets associated with the facet; and

a computer system having one or more software pieces each having a plurality of lines of computer instructions wherein the computer instructions are executed by the computer system, the software pieces further comprising a query engine that expands the one or more query terms of the query using the one or more synsets and the one or more taxonomies in the knowledge base to generate an expanded query, selects an interpretation of a concept from the expanded query and the corpus of documents, selects one or more facets that match the interpretation of the expanded query based on the corpus of documents and performs a deep concept query using the interpretation of the expanded query and the selected one or more facets against the corpus of documents.

2. The system of claim 1, wherein the query engine further comprises a synset expansion engine that, when a particular query term matches a particular term in a particular synset, expands the query based on the particular synset and retrieves one or more documents from the corpus of documents having at least one of the group of terms contained in the synset.

3. The system of claim 1, wherein the query engine further comprises a facet display engine that, when a particular synset associated with a particular facet matches a particular query term, triggers the particular facet and displays the particular facet.

4. The system of claim 3, wherein the facet display engine further comprises a graphical facet display engine that graphically displays the one or more documents associated with the particular facet.

5. The system of the claim 1, wherein the query engine further comprises a double dip engine that performs a second query based on the synset expansion and the taxonomy expansion of the one or more query terms.

6. The system of claim 1, wherein the query engine further comprises a localization engine that retrieves one or more documents from the corpus of documents that are localized to a localization parameter selected by the user.

7. The system of claim 6, wherein the localization parameter further comprises a neighborhood, a vicinity, a town or a county associated with a geographical term.

8. The system of claim 1, wherein the query engine further comprises a Boolean term parsing engine that parses a query containing one or more Boolean terms in the query.

9. The system of claim 8, wherein the Boolean terms further comprises AND, OR, NOT and ANDNOT.

10. The system of claim 1, wherein the formally represented knowledge comprises one or more of a declarative entity list, a topic map, a thesaurus, and a semantic network.

24. The system of claim 3 further comprising a facet ontology engine that, when a particular term in the query matches a particular synset in the facet, performs a query using the particular facet.

11. A computer implemented query processing method for processing a query having one or more query terms using formally expressed knowledge against a corpus of documents based on a knowledge base containing a plurality of pieces of formally expressed knowledge, wherein each piece of formally represented knowledge further comprises an item that has been edited or analyzed, the formally represented knowledge further comprising one or more synsets wherein each synset contains a group of terms that have a same meaning, one or more taxonomies wherein each taxonomy contains one or more synsets in a subject matter area that are organized from a synset having a general meaning to a synset having a specific meaning, one or more ontologies wherein each ontology contains one or more synsets associated with an area of interest and one or more facets wherein each facet is associated with a particular ontology and wherein a document is associated with the facet when the document contains the one or more synsets associated with the facet, the method comprising:

expanding, using a hardware query engine, the one or more query terms of the query using the one or more synsets and the one or more taxonomies in the knowledge base to generate an expanded query;

selecting an interpretation of a concept from the expanded query and the corpus of documents;

selecting one or more facets that match the interpretation of the expanded query based on the corpus of documents; and

performing, using the hardware query engine, a deep concept query using the expanded query and the selected one or more facets against a corpus of documents.

12. The method of claim 11, wherein expanding the query terms further comprises expanding, when a particular query term matches a particular term in a particular synset, the query based on the particular synset and retrieving one or more documents from the corpus of documents having at least one of the group of terms contained in the synset.

13. The method of claim 11, wherein expanding the query terms further comprises triggering a particular facet and displaying the particular facet.

14. The method of claim 13, wherein displaying the particular facet further comprises graphically displaying the one or more documents associated with the particular facet.

15. The method of the claim 11, wherein performing the deep concept query further comprises performing a second query based on the synset expansion and the taxonomy expansion of the one or more query terms.

16. The method of claim 11, further comprising retrieving one or more documents that are localized to a localization parameter selected by the user.

17. The method of claim 16, wherein the localization parameter further comprises a neighborhood, a vicinity, a town or a county associated with a geographical term.

18. The method of claim 11 further comprising retrieving a set of documents that further comprises parsing a query containing one or more Boolean terms in the query.

19. The method of claim 18, wherein the Boolean terms further comprises AND, OR, NOT and AND NOT.

20. The method of claim 11 wherein the formally expressed knowledge comprises one or more of a declarative entity list, a topic map, a thesaurus, and a semantic network.

25. The method of claim 13 further comprising performing a second query, when a particular term in the query matches a particular synset in the facet, using the facet.

21. A computer-implemented query processing method in which the query is interpreted based on a set of formally represented knowledge items stored in a knowledge base, wherein each formally represented knowledge item further comprises an item that has been edited or analyzed, the formally represented knowledge further comprising one or more synsets wherein each synset contains a group of terms that have a same meaning, one or more taxonomies wherein each taxonomy contains one or more synsets in a subject matter area that are organized from a synset having a general meaning to a synset having a specific meaning, one or more ontologies wherein each ontology contains one or more synsets associated with an area of interest and one or more facets wherein each facet is associated with a particular ontology and wherein a piece of content is associated with the facet when the piece of content contains the one or more synsets associated with the facet, the method comprising:

receiving a query containing one or more query terms;

determining a language of the query;

identifying one or more items in the knowledge base based on the language of the query that match the one or more terms in the query;

performing a raw search of the one or more terms of the query against a corpus of pieces of content to retrieve a raw search results set containing one or more pieces of content with a term that matches a term of the query;

selecting an interpretation of a concept based on the query and the retrieved pieces of content;

modifying the query to generate an expanded query that includes the identified knowledge base items if the raw search results set match one or more of the identified knowledge base items;

selecting one or more facets that match the interpretation of the expanded query based on the corpus of pieces of content; and

performing a concept search that compares the one or more query terms against the expanded query and the selected one or more facets against a corpus of documents to display one or more facets whose synsets are matched by the query terms.

22. The method of claim 21 wherein the formally represented knowledge comprises one or more of a declarative entity list, a topic map, a thesaurus and a semantic network.

23. The method of claim 21, wherein determining the language of the query further comprises one of determining the language of the query based on an explicit user preference, and determining the language of the query based on comparing the languages of pieces of content that contain the query terms.