Cluster-Based Ranking with a Behavioral Web Graph

• US 20080313171A1
• Filed: 12/06/2007
• Published: 12/18/2008
• Est. Priority Date: 06/12/2007
• Status: Abandoned Application

• Current Assignee: Tethys Prime, Inc.
• Original Assignee: Tethys Prime, Inc.
• Inventors: Galvin, Brian
• Application Number:
  US11/951,826
• Publication Number:
  US 20080313171A1
• Time in Patent Office: Days
• US Class Current: 707/5
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques

Abstract

A computer implemented method for returning relevant nodes in a network search has steps for (a) creating a behavioral network graph having points relating pairs of network nodes with values at the points indicating probability that a user connected at one node of the pair will transition next to the other node of the pair; (b) determining node clusters based on relatively high probability of transition between nodes in the cluster; (c) entering search criteria for finding nodes, and noting and returning nodes that satisfy the search criteria; and (d) returning in addition nodes in one or more clusters associated with one or more nodes that satisfy the search criteria.

Claims

1. A computer implemented method for returning relevant nodes in a network search, comprising the steps of:

(a) creating a behavioral network graph having points relating pairs of network nodes with values at the points indicating probability that a user connected at one node of the pair will transition next to the other node of the pair;

(b) determining node clusters based on relatively high probability of transition between nodes in the cluster;

(c) entering search criteria for finding nodes, and noting and returning nodes that satisfy the search criteria; and

(d) returning in addition nodes in one or more clusters associated with one or more nodes that satisfy the search criteria.

2. The method of claim 1 further comprising a step for ranking nodes returned based on relative probability of transition between the nodes.

3. The method of claim 1 wherein the network is the Internet network, and nodes are pages in the Internet network.

4. The method of claim 1 wherein the behavioral network graph is a square matrix with a unique row and a unique column associated with each node, points in the map being intersections of rows and columns in the matrix, each intersection defining a probability of transition between the two associated nodes at the intersection.

5. The method of claim 4 wherein the network graph is a web graph, and the nodes are web pages in the Internet.

6. A computerized system for returning relevant nodes in a network search, comprising:

a behavioral network graph having points relating pairs of network nodes with values at the points indicating probability that a user connected at one node of the pair will transition next to the other node of the pair;

a mechanism for determining node clusters based on relatively high probability of transition between nodes in the cluster;

a search engine for entering search criteria for finding nodes, and for noting and returning nodes that satisfy the search criteria; and

a mechanism for returning in addition, nodes in one or more clusters associated with one or more nodes that satisfy the search criteria.

7. The system of claim 6 further comprising a step for ranking nodes returned based on relative probability of transition between the nodes.

8. The system of claim 6 wherein the network is the Internet network, and nodes are pages in the Internet network.

9. The system of claim 6 wherein the behavioral network graph is a square matrix with a unique row and a unique column associated with each node, points in the map being intersections of rows and columns in the matrix, each intersection defining a probability of transition between the two associated nodes at the intersection.

10. The system of claim 9 wherein the network graph is a web graph, and the nodes are web pages in the Internet.

11. A method for relating persons by interest, comprising steps of:

(a) creating a behavioral network graph having points relating pairs of network nodes with values at the points indicating probability that a user connected at one node of the pair will transition next to the other node of the pair;

(b) determining node clusters based on relatively high probability of transition between nodes in the cluster; and

(c) using node identifiers in determined clusters for directing information to the nodes in the clusters.

12. The method of claim 11 wherein the directed information is advertising information.

13. The method of claim 11 wherein the network is a communication network and nodes are addresses of communication devices associated with specific users.

14. The method of claim 13 including a step for determining an interest or characteristic associated with one of the specific users, and assigning that interest or characteristic to one or more other specific users identified by determining node clusters.