SYSTEM AND PROGRAM FOR HANDLING ANCHOR TEXT

• US 20090083270A1
• Filed: 12/03/2008
• Published: 03/26/2009
• Est. Priority Date: 01/26/2004
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Kraft, Reiner, Neumann, Andreas
• Granted Patent: US 8,285,724 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9558 Details of hyperlinks; Management of linked annotations
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access

Abstract

Disclosed is a system and program for processing anchor text for information retrieval. A set of anchors that point to a target document is formed. Anchors with same anchor text are grouped together. Information is computed for each group. Context information is generated for the target document based on the computed information.

Claims

1. A computer system, comprising:

a processor;

storage; and

hardware logic for;

forming a set of anchors that point to a target document, wherein each anchor is a path from a source document to the target document;

grouping together anchors with same anchor text, wherein each anchor is associated with anchor text;

computing a relevance score for each group, wherein computing the relevance score includes computing a linguistic score for each group; and

generating context information for the target document based on the computed relevance score, wherein a title is composed from text of a group with a highest relevance score and a summary of the target document is composed from anchor texts of a number of groups with highest relevance scores.

2. The computer system of claim 1, wherein the logic further comprises:

determining a language of each document in a collection of documents;

determining rank of each document in the collection of documents; and

determining a proximity class of each document in the collection of documents, wherein the proximity class specifies how close a source document is to the target document.

3. The computer system of claim 1, wherein the logic further comprises:

determining a predominant language in the set of anchors; and

pruning anchors from the set that are not in the predominant language.

4. The computer system of claim 1, wherein the logic further comprises:

pruning anchors from the set that include at least a portion of a path to the target document.

5. The computer system of claim 1, wherein the logic further comprises:

pruning anchors based on a configurable set of words.

6. The computer system of claim 1, wherein the logic for computing the relevance score further comprises:

computing a weighted sum of occurrences for anchor text for anchors in each group, wherein a weight of each individual occurrence of the anchor text is determined by a proximity class of an anchor and a weight associated with that proximity class.

7. The computer system of claim 1, wherein the logic for computing the relevance score further comprises:

computing an accumulated rank for each group.

8. Them computer system of claim 1, wherein the logic for computing the relevance score further comprises:

generating the relevance score for each group based on a weighted sum of occurrences, an accumulated rank, and a linguistic score.

9. An article of manufacture comprising a computer readable medium including a program for processing anchor text in documents, the program when executed by a processor causes operations to be performed, the operations comprising:

forming a set of anchors that point to a target document, wherein each anchor is a path from a source document to the target document;

grouping together anchors with same anchor text, wherein each anchor is associated with anchor text;

computing a relevance score for each group, wherein computing the relevance score includes computing a linguistic score for each group; and

generating context information for the target document based on the computed relevance score, wherein a title is composed from text of a group with a highest relevance score and a summary of the target document is composed from anchor texts of a number of groups with highest relevance scores.

10. The article of manufacture of claim 9, wherein the operations further comprise:

determining a language of each document in a collection of documents;

determining rank of each document in the collection of documents; and

determining a proximity class of each document in the collection of documents, wherein the proximity class specifies how close a source document is to the target document.

11. The article of manufacture of claim 9, wherein the operations further comprise:

determining a predominant language in the set of anchors; and

pruning anchors from the set that are not in the predominant language.

12. The article of manufacture of claim 9, wherein the operations further comprise:

pruning anchors from the set that include at least a portion of a path to the target document.

13. The article of manufacture of claim 9, wherein the operations further comprise:

pruning anchors based on a configurable set of words.

14. The article of manufacture of claim 9, wherein the operations for computing the relevance score further comprise:

computing a weighted sum of occurrences for anchor text for anchors in each group, wherein a weight of each individual occurrence of the anchor text is determined by a proximity class of an anchor and a weight associated with that proximity class.

15. The article of manufacture of claim 9, wherein the operations for computing the relevance score further comprise:

computing an accumulated rank for each group.

16. The article of manufacture of claim 9, wherein the operations for computing the relevance score further comprise:

generating the relevance score for each group based on a weighted sum of occurrences, an accumulated rank, and a linguistic score.