Method and system for mining information based on relationships

• US 7,529,735 B2
• Filed: 02/11/2005
• Issued: 05/05/2009
• Est. Priority Date: 02/11/2005
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Ma, Wei-Ying, Chen, Zheng, Zhang, Benyu, Zeng, Hua-Jun, Xu, Gu, Gao, Hongbin, Hinrichs, Randy
• Primary Examiner(s): Boccio; Vincent F
• Assistant Examiner(s): Hotelling; Harold A
• Application Number:
  US11/057,100
• Publication Number:
  US 20060184481A1
• Time in Patent Office: 1,544 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 16/313 Selection or weighting of terms for indexing
• CPC Class Codes: G06N 5/022 Knowledge engineering; Knowledge acquisition
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99934 Query formulation, input preparation, or translation
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access

Abstract

A method and system for identifying information about people is provided. The information system identifies groups of people that have relationships based on their relationships to documents or more generally to objects. The information system initially is provided with an indication of which people have which relationships to which documents. The information system then identifies clusters of people based on having a relationship to the same objects. The information system may also identify clusters of related objects associated with a cluster of people. When a user wants to identify information about a person, the user can provide the name of that person to the information system. The information system then can retrieve and display the names of the other people who are in the same cluster as the person.

Claims

1. A computer-implemented method for a computer to identify authors that have a relationship, the method comprising:

providing an indication of authors of papers;

identifying clusters of authors based on co-author relationships between the authors wherein authors within a same cluster have a relationship;

storing in a memory of the computer an indication of the identified clusters of authors; and

calculating an importance of authors within an author cluster

wherein the importance of an author is defined recursively based on the importance of the papers authored by the author and

wherein the importance of an author and a paper is defined as follows;

R_author=W diag(R_paperN^T_paper)

R_paper=W^T diag(R_authorN^T_author)

where R_author is a vector of the importance (or ranking) of the authors, R_paper is a vector of the importance (or ranking) of the papers, W is an adjacency matrix mapping authors to papers, W^T is a transpose of the adjacency matrix W, N^T_paper is a matrix of normalization terms for the papers, and N^T_author is a matrix of normalization terms for the authors.

2. The computer-implemented method of claim 1 wherein the identifying of author clusters includes:

generating a publication vector for each author, a publication vector identifying papers authored by the author; and

clustering the authors based on distance between the publication vectors.

3. The computer-implemented method of claim 2 wherein the distance is defined as follows:

Dist(A,B)=1−

inner(A,B)/min(|A|2,|B|2)

where inner( )is the dot product and |.| is the norm.

4. The computer-implemented method of claim 1 including clustering papers of authors within an author cluster into clusters of topics.

5. The computer-implemented method of claim 4 including identifying descriptions of a topic of the papers of a paper cluster.

6. The computer-implemented method of claim 5 wherein the identifying includes ranking phrases within the papers of a topic based on descriptiveness of the topic.

7. The computer-implemented method of claim 6 wherein the ranking of phrases includes:

ranking two-word phrases within the papers;

selecting high ranking two-word phrases; and

repeatedly extending the selected phrases based on the ranking of the extended phrases.

8. The computer-implemented method of claim 4 including identifying information relating to an author by receiving an indication of an author and a topic and extracting information about the author from papers authored by that author within that topic.

9. The computer-implemented method of claim 4 including identifying an author as a rising star within the topic of the paper cluster based on increasing importance of that author to that topic over time.

10. The computer-implemented method of claim 4 including identifying an influential author to an author based on importance of the authors within each paper cluster for which the author authored a paper.

11. A computer-readable storage medium containing instructions for controlling a computer system to provide information about authors that have a relationship, by a method comprising:

accessing indications of authors of papers to identify clusters of authors based on co-author relationships between the authors wherein authors within a same cluster have a co-author relationship;

storing an indication of the identified clusters of authors; and

calculating an importance of authors within an author cluster

wherein the importance of an author is defined recursively based on the importance of the papers authored by the author and

wherein the importance of an author and a paper is defined as follows;

R_author=W diag(R_paperN^T_paper)

R_paper=W^T diag(R_authorN^T_author)

where R_author is a vector of the importance of the authors, R_paper is a vector of the importance of the papers, W is an adjacency matrix mapping authors to papers, W^T is a transpose of the adjacency matrix W, N^T_paper is a matrix of normalization terms for the papers, and N^T_author is a matrix of normalization terms for the authors.

12. The computer-readable storage medium of claim 11 wherein the identifying of author clusters includes:

generating a publication vector for each author, a publication vector identifying papers authored by the author; and

clustering the authors based on distance between the publication vectors.

13. The computer-readable storage medium of claim 12 wherein the distance is defined as follows:

Dist(A,B)=1−

inner(A,B)/min(|A|2,|B|2)

where inner( )is the dot product and |.| is the norm.

14. The computer-readable storage medium of claim 11 including clustering papers of authors within an author cluster into clusters of topics.

15. The computer-readable storage medium of claim 14 including identifying descriptions of a topic of the papers of a paper cluster.

16. The computer-readable storage medium of claim 15 wherein the identifying includes ranking phrases within the papers of a topic based on descriptiveness of the topic.

17. The computer-readable storage medium of claim 16 wherein the ranking of phrases includes:

ranking two-word phrases within the papers;

selecting high ranking two-word phrases; and

repeatedly extending the selected phrases based on the ranking of the extended phrases.

18. The computer-readable storage medium of claim 14 including identifying information relating to an author by receiving an indication of an author and a topic and extracting information about the author from papers authored by that author within that topic.

19. The computer-readable storage medium of claim 14 including identifying an author as a rising star within the topic of the paper cluster based on increasing importance of that author to that topic over time.

20. The computer-readable storage medium of claim 14 including identifying an influential author to an author based on importance of the authors within each paper cluster for which the author authored a paper.