Semantic network clustering influenced by index omissions

• US 20110289039A1
• Filed: 08/02/2011
• Published: 11/24/2011
• Est. Priority Date: 11/03/2009
• Status: Active Grant

• Current Assignee: Postq IPR OY
• Original Assignee: TATU Ylonen OY Limited
• Inventors: Ylonen, Tatu J.
• Granted Patent: US 8,666,923 B2
• Time in Patent Office: Days
• US Class Current: 706/47
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation

Abstract

Some links are omitted from indexes associated with semantic network elements and clustering is influenced by such omissions.

Claims

1. A computer comprising:

a semantic network (104) comprising a first semantic network element (105) associated with a first index (113) and a link (106) that refers to the first semantic network element;

an index maintainer (107,115) coupled to at least one omission rule (108,116) configured to determine for the link whether the link is to be included in an index associated with a semantic network element that the link references; and

a clusterer (111) configured to cause related objects to be clustered near each other in an address space of the computer;

wherein the clusterer causes the link to be clustered with the first semantic network element in response to the index maintainer adding the link to the first index and omitting the link from an index associated with another node that the link refers to.

2. The computer of claim 1, wherein the semantic network elements are nodes in a semantic network.

3. The computer of claim 1, wherein the omission rule is responsive to the size of the index.

4. The computer of claim 1, wherein the omission rule is responsive to the index being lossy.

5. The computer of claim 1, wherein the omission rule is responsive to the index being associated with a semantic network element being lossy.

6. A method comprising:

adding (206), by a computer, a link (106) referring to a first node (105) and a second node (105) in a semantic network (104), to an index (113) associated with the first node;

omitting (204) the link from an index associated with the second node; and

clustering the link with the first node.

7. The method of claim 6, wherein the omitting is responsive to the size of the index.

8. The method of claim 6, wherein the omitting is responsive to the index being lossy.

9. The method of claim 6, wherein the omitting is responsive to the index being associated with a node that is lossy.

10. A computer program product stored on a non-transitory computer-readable medium, comprising:

computer readable program code means for associating at least one index with each of a plurality of nodes in a semantic network;

computer readable program code means for adding a link to an index associated with one of the nodes the link refers to;

computer readable program code means for omitting the link from an index associated with another one of the nodes the link refers to; and

computer readable program code means for clustering the link with a node it refers to and from whose index the link was not omitted.

11. The computer program product of claim 10, wherein the computer program code means for omitting is responsive to the size of the index.

12. The computer program product of claim 10, wherein the computer program code means for omitting is responsive to the index being lossy.

13. The computer program product of claim 10, wherein the computer program code means for omitting is responsive to the index being associated with a network element that is lossy.