# US 2010/0076910 A1 · Calculating web page importance based on web behavior model

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2010/0076910 A1
- Document ID: US20100076910A1
- Status: Published application
- Kind: A1
- Application number: 12/237,392
- Filed: 2008-09-25
- Published: 2010-03-25
- Earliest claimed: 2008-09-25
- Assignees: Microsoft Corporation
- Inventors: Bin Gao; Tie-Yan Liu; Hang Li; Yuting Liu
- Web page: https://patentgrep.com/patent/US20100076910A1

## CPC

- G06Q 30/02: Marketing; Price estimation or determination; Fundraising
- G06F 16/951: Indexing; Web crawling techniques
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Method for determining a webpage importance, including receiving web browsing behavior data of one or more users; creating a model of the web browsing behavior data; calculating a stationary probability distribution of the model; and correlating the stationary probability distribution to the webpage importance.

## Claims

### Claim 1

A method for determining a webpage importance, comprising:
receiving web browsing behavior data of one or more users;
creating a model of the web browsing behavior data;
calculating a stationary probability distribution of the model; and
correlating the stationary probability distribution to the webpage importance.

### Claim 2 (depends on 1)

The method of claim 1, wherein the web browsing behavior data comprises one or more web pages accessed by the users, a record of times at which the users accessed each web page, and a method in which each web page is accessed.

### Claim 3 (depends on 1)

The method of claim 1, further comprising:
creating a browsing graph representing the web pages accessed by the users and the method in which each web page is accessed.

### Claim 4 (depends on 3)

The method of claim 3, wherein the browsing graph is created by:
organizing the web browsing behavior data into a browsing thread;
adding a node on the browsing graph for each web page listed on the browsing thread; and
adding a line connecting two or more nodes on the browsing graph, wherein the line represents a hyperlink for accessing a webpage.

### Claim 5 (depends on 4)

The method of claim 4, wherein the browsing thread separates the web browsing behavior data for each user.

### Claim 6 (depends on 1)

The method of claim 1, wherein the model comprises a continuous-time time-homogeneous Markov process model.

### Claim 7 (depends on 1)

The method of claim 1, wherein creating the model comprises:
estimating a time in which each web page was being accessed;
calculating a reset probability in which a web user navigated from a first webpage to a second webpage by entering a Uniform Resource Locator (URL) address into a web browser;
determining a transition probability in which the web user navigated from the first webpage to the second webpage by clicking a hyperlink on the first webpage;
using the reset probability and the transition probability to create a discrete-time embedded Markov chain model; and
using the time and the discrete-time embedded Markov chain model to create the model.

### Claim 8 (depends on 7)

The method of claim 7, wherein estimating the time comprises:
determining a difference in time at which the first webpage was accessed and a time at which the second webpage was accessed; and
removing noise from the time.

### Claim 9 (depends on 8)

The method of claim 8, wherein removing the noise comprises:
solving an optimization problem according to
min
q
ii

(
(
Z
_
+
1
q
ii
)
-
1
2

(
S
2
-
1
q
ii
2
)
)
2
,

s
.
t
.

q
ii
<
0
wherein a mean and variance from samples of a random variable Z are Z-bar ({tilde over (Z)}) and S2 and the time in which the URL address was being accessed is qii.

### Claim 10 (depends on 7)

The method of claim 7, further comprising:
segmenting a web user's activities into one or more sessions; and
calculating the time in which each web page was being accessed, the reset probability and the transition probability using the sessions.

### Claim 11 (depends on 10)

The method of claim 10, wherein segmenting the web user's activities into sessions comprises:
determining whether the URL address was entered into the web browser; and
creating a new session if the URL address has been entered.

### Claim 12 (depends on 10)

The method of claim 10, wherein segmenting the web user's activities into sessions comprises:
determining a difference in time at which the first webpage was accessed and the time at which the second webpage was accessed;
comparing the difference in time to a predetermined value; and
creating a new session if the difference in time is greater than the predetermined value.

### Claim 13 (depends on 10)

The method of claim 10, further comprising: pairing two URL addresses if the two URL addresses were accessed in sequence.

### Claim 14

A method for creating a browsing graph representing web browsing behavior data, comprising:
receiving the web browsing behavior data from one or more web browsers;
organizing the web browsing behavior data into a browsing thread;
adding a node on the browsing graph for each web page listed on the browsing thread;
adding a line connecting two or more nodes, the line representing a hyperlink for accessing a webpage; and
adding one or more times in which each webpage was accessed in a metadata of the browsing graph.

### Claim 15 (depends on 14)

The method of claim 14, wherein the browsing thread comprises:
a machine identification number;
a list of web pages that have been accessed;
a time in which the web pages have been accessed; and
a method in which each web page was accessed.

### Claim 16 (depends on 15)

The method of claim 15, wherein the web pages are listed in chronological order.

### Claim 17 (depends on 14)

The method of claim 14, wherein the line is an arrow representing accessing a second web page from a first web page.

### Claim 18

A computer-readable medium having stored thereon computer-executable instructions which, when executed by a computer, cause the computer to:
receive web browsing behavior data of one or more users;
organize the web browsing behavior data into a browsing thread;
add a node on a browsing graph for each web page listed on the browsing thread;
add a line connecting two or more nodes on the browsing graph, the line representing a hyperlink for accessing a webpage;
add one or more times in which each webpage was accessed in a metadata of the browsing graph;
create a model of the browsing graph;
calculate a stationary probability distribution of the model; and
correlate the stationary probability distribution to a webpage importance.

### Claim 19 (depends on 18)

The computer-readable medium of claim 18, wherein the model is created by:
estimating a time in which each web page was being accessed;
calculating a reset probability in which the users navigated from a first webpage to a second webpage by entering a Uniform Resource Locator (URL) address into a web browser;
determining a transition probability in which the users navigated from the first webpage to the second webpage by clicking a hyperlink on the first webpage;
using the reset probability and the transition probability to create a discrete-time embedded Markov chain model; and
using the time and the discrete-time embedded Markov chain model to create the model.

### Claim 20 (depends on 19)

The computer-readable medium of claim 19, wherein estimating the time in which each web page was being accessed comprises:
determining a difference in time at which the first webpage was accessed and the time at which the second webpage was accessed; and
removing noise from the time.
