Method and apparatus for semantic characterization of general content streams and repositories

• US 6,108,619 A
• Filed: 07/02/1998
• Issued: 08/22/2000
• Est. Priority Date: 07/02/1998
• Status: Expired due to Term

• Current Assignee: RPX Corporation
• Original Assignee: Novell Incorporated (Open Text Corporation)
• Inventors: Carter, Stephen R., LaVange, Jr., Donald H., Jensen, Delos C.
• Primary Examiner(s): Thomas, Joseph
• Application Number:
  US09/109,804
• Time in Patent Office: 782 Days
• Field of Search: 704/9, 704/10, 704/8, 704/1, 707/530, 707/531, 707/532, 707/533, 707/1, 707/2, 707/3, 707/4, 707/5, 707/6, 707/100, 707/102, 707/103, 707/104
• US Class Current: 704/9
• CPC Class Codes: G06F 40/268 Morphological analysis
• CPC Class Codes: G06F 40/289 Phrasal analysis, e.g. finite state techniques or chunking
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

A method in a system comprises accessing a content stream and parsing the content stream into phrases. One or more tokens are determined, at least partially based on the extracted phrases. One or more profiles are associated with the tokens. One or more semantic records are instantiated from the profiles. The semantic record is compared with other semantic records.

Claims

1. A system for assigning semantic characterization to a content stream, comprising:

a) a network having a plurality of principals;

b) the content stream in the network associated with at least one principal, said content stream having a plurality of phrases;

c) a plurality of tokens each associated with one or more phrases, each token comprising metadata about one or more of the associated phrases;

d) a plurality of profiles each associated with one or more tokens; and

e) one or more semantic records each associated with a profile and a principal.

2. A system as recited in claim 1, further comprising a token dictionary having associations between the tokens and the phrases.

3. A system as recited in claim 1, further comprising a profile store having associations between the profiles and its corresponding tokens.

4. A system as recited in claim 1, wherein each profile defines a semantic concept.

5. A system as recited in claim 1, wherein each profile defines a topological vector space corresponding to the associated semantic record.

6. A system as recited in claim 5, wherein the profile further defines a probability density function within the topological vector space.

7. A system as recited in claim 1, further comprising a means for comparing one or more semantic records.

8. A system as recited in claim 1, further comprising a means for managing the tokens and profiles.

9. A method in a computer system for assigning semantic characterization to a content stream, comprising the steps of:

a) accessing the content stream;

b) extracting tokens from the content stream, each token comprising metadata about a part of the content stream;

c) associating one or more profiles with the extracted tokens; and

d) representing at least a portion of the content stream in a semantic space corresponding to one or more profiles.

10. A method as recited in claim 9, wherein the step of extracting comprises:

parsing the content stream into phrases; and

abstracting metadata about the phrases.

11. A method as recited in claim 10, wherein the step of abstracting comprises the steps of:

a) accessing a token dictionary having associations between the tokens and the phrases; and

b) determining tokens from the token dictionary based on the parsed phrases.

12. A method as recited in claim 9, wherein the step of representing comprises instantiating a profile.

13. A method as recited in claim 12, wherein the profile is instantiated only if a token threshold is satisfied.

14. A method as recited in claim 9, further comprising the step of comparing at least a portion of the semantic space with one or more portions of other semantic spaces.

15. A method as recited in claim 14, wherein the step of comparing comprises normalizing the semantic spaces.

16. A method as recited in claim 9, further comprising the step of creating a second semantic space based on the semantic space.

17. A method as recited in claim 9, further comprising the step of querying the semantic space.

18. A method as recited in claim 9, wherein the step of representing occurs only if a frequency threshold of tokens is satisfied.

19. A computer readable medium for assigning semantic characterization to a content stream, comprising instructions capable of performing the method of claim 9.

20. A method in a computer system for assigning semantic characterization to a content stream, comprising the steps of:

a) accessing a content stream;

b) parsing the content stream into phrases;

c) determining one or more tokens at least partially based on the phrases parsed from the content stream, each token comprising metadata about one or more of the associated phrases;

d) associating one or more profiles with the tokens;

e) instantiating a semantic record from the profiles; and

f) comparing the semantic record with other semantic records.