Information retrieval

• US 8,069,185 B2
• Filed: 02/24/2004
• Issued: 11/29/2011
• Est. Priority Date: 03/25/2003
• Status: Active Grant

• Current Assignee: British Telecommunications PLC (BT Group PLC)
• Original Assignee: British Telecommunications PLC (BT Group PLC)
• Inventors: Azvine, Behnam, Martin, Trevor P
• Primary Examiner(s): BROWN, SHEREE N
• Application Number:
  US10/549,365
• Publication Number:
  US 20060179074A1
• Time in Patent Office: 2,835 Days
• Field of Search: 707 1- 3
• US Class Current: 707/802
• CPC Class Codes: G06F 16/3338 Query expansion
• CPC Class Codes: G06F 16/374 Thesaurus

Abstract

A concept dictionary is generated and/or updated in respect of an information system. The concept dictionary is used to assist in selecting queries and query terms for use in interrogating the information system. A lexical reference source is first used to generate queries semantically related to a query entered by a user, and the answers returned for each query are analyzed using a fuzzy processing technique to determine semantic relationships between the queries. The queries and the determined relationships are recorded in a concept dictionary for subsequent use.

Claims

1. A method of generating a concept dictionary for use in querying an information system comprising:

(i) receiving an information search criterion;

(ii) deriving from said received search criterion, using a lexical reference source, at least one different search criterion having related meaning to said received search criterion;

(iii) identifying a set of information in said information system relevant to said received search criterion and a different set of information in said information system relevant to said at least one derived search criterion;

(iv) analyzing the identified sets of information and deriving from similarities and differences therebetween relationships between said received search criterion and said at least one derived search criterion in the context of said information system; and

(v) storing, in a concept dictionary, information relating to said received and said at least one derived search criterion and to respective said derived relationships therebetween, for use in querying said information system.

2. A method as in claim 1, wherein, at step (i), receiving an information search criterion comprises selecting an information search criterion stored in said concept dictionary.

3. A method as in claim 1, wherein, at step (ii), deriving at least one search criterion having related meaning comprises replacing a term of said received search criterion with a related term having a more specific meaning according to said lexical reference source.

4. A method as in claim 1, wherein, at step (ii) deriving at least one search criterion having related meaning comprises replacing a term of said received search criterion with a related term having a more general meaning according to said lexical reference source.

5. A method as in claim 1, wherein, at step (ii) deriving at least one search criterion having related meaning comprises replacing a term of said received search criterion with a related term having an equivalent meaning according to said lexical reference source.

6. A method as in claim 1, wherein, at step (ii), said lexical reference source is a thesaurus.

7. A method as in claim 1, wherein, at step (ii), said lexical reference source is an ontological database.

8. A method as in claim 1, wherein, at step (ii), a plurality of search criteria are derived, each having related meaning to said received search criterion, and wherein at step (iv), the respective identified sets of information are analyzed to derive relationships between search criteria comprised in said plurality of derived search criteria.

9. A method as in claim 1, wherein, at step (iv), deriving relationships between said search criteria comprises performing fuzzy processing of said derived search criteria and respective said identified sets of information to determine a measure of the and/or of one said search criterion over another in the context of said information system.

10. A method of accessing sets of information stored in an information system using information search criteria stored in a concept dictionary generated for the information system according to the method in claim 1, further comprising:

(a) selecting a first information search criterion;

(b) using a search engine to identify one or more sets of information in the information system relevant to said first information search criterion; and

(c) selecting at least one further information search criterion from search criteria stored in the concept dictionary, semantically related to said first information search criterion according to information stored in the concept dictionary, according to whether a more general, a more specialized or an equivalent search is required.

11. An information retrieval apparatus for accessing sets of information stored in an information system, said apparatus comprising;

an input for receiving an information search criterion;

deriving means for deriving from said received search criterion, using a lexical reference source, at least one different search criterion having related meaning to said received information search criterion;

retrieval means for identifying a set of information in said information system relevant to said received search criterion and a different set of information in said information system relevant to said at least one derived search criterion;

analysis means (135) for an analyzing said identified sets of information and deriving from similarities and differences therebetween relationships between said received search criterion and said at least one derived search criterion in the context of said information system; and

updating means for storing, in a concept dictionary, information relating to said received and said at least one derived search criterion and to respective said derived relationships therebetween, for use in querying said information system.