# US 2014/0195207 A1 · Estimating probability of spreading information by users on micro-weblogs

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

- Number: US 2014/0195207 A1
- Document ID: US20140195207A1
- Status: Published application
- Kind: A1
- Application number: 13/734,556
- Filed: 2013-01-04
- Published: 2014-07-10
- Earliest claimed: 2013-01-04
- Assignees: International Business Machines Corporation
- Inventors: Miki Enoki; Akiko Suzuki; Toyotaro Suzumura; Michiaki Tatsubori
- Web page: https://patentgrep.com/patent/US20140195207A1

## CPC

- G06Q 30/02: Marketing; Price estimation or determination; Fundraising
- G06Q 10/42: Determination of affinities or common interests between users
- G06Q 10/48: using social graphs
- 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

## Abstract

Methods and systems for estimating a probability of re-sharing information include extracting keywords from a set of documents addressed to a user; weighting the keywords from the set of documents according a metric for the user's interest in the keywords' respective source documents to create an interest model; receiving a new document having one or more keywords; and determining a likelihood that the user will re-share the new document, where the likelihood is based on the interest model and the one or more keywords present in the new document.

## Claims

### Claim 1

A method of estimating a probability of re-sharing information, comprising:
extracting keywords from a set of documents addressed to a user;
weighting the keywords from the set of documents according a metric for the user's interest in the keywords' respective source documents to create an interest model;
receiving a new document having one or more keywords; and
determining a likelihood that the user will re-share the new document using a processor, said likelihood being based on the interest model and the one or more keywords present in the new document.

### Claim 2 (depends on 1)

The method of claim 1, weighting a keyword comprises calculating a weight based on a proportion of documents including the keyword that were re-shared by the user.

### Claim 3 (depends on 2)

The method of claim 2, wherein weighting a keyword comprises increasing the weight of a re-shared keyword in inverse proportion to a level of activity for the user.

### Claim 4 (depends on 2)

The method of claim 2, wherein weighting a keyword comprises calculating a weight according to:
q
j
(
u
)
=
∑
m


r
(
m
)

f
j

(
m
)
∑
m


a

(
t
m
)

f
j

(
m
)
+
C
,
where r(m) is 1 if a post m was re-shared by the user u and 0 if the post m is ignored, where fj(m) is 1 if the keyword j appeared in a post m and is 0 otherwise, and where C is a small constant.

### Claim 5 (depends on 1)

The method of claim 1, wherein determining a likelihood comprises calculating a probability that the user will re-share the new document as:
P
(
u
,
m
)
=
a

(
t
m
)

(
1
-
∏
j


(
1
-
q
j
(
u
)

f
j

(
m
)
)
)
,
where fj(m) is 1 if the keyword j appeared in a post m and is 0 otherwise and where qj(u) is a weighted value for the keyword j for the user u.

### Claim 6 (depends on 1)

The method of claim 1, wherein the keywords match a set of specific predetermined keywords.

### Claim 7 (depends on 1)

The method of claim 1, wherein the keywords include at least one web address.

### Claim 8 (depends on 1)

A non-transitory computer readable storage medium comprising a computer readable program for estimating a probability of re-sharing information, wherein the computer readable program when executed on a computer causes the computer to perform the steps of claim 1.

### Claim 9

A method of predicting the spread of information, comprising:
determining a current spread of information comprising source users who have shared a target document and edge users who have received the target document;
for each edge user, determining a likelihood that the user will re-share the target document, by:
extracting keywords from a historical set of documents addressed to the user;
weighting the extracted keywords from the historical set of documents according a metric for the user's interest in the extracted keywords' respective source documents to create an interest model; and
determining a likelihood that the user will re-share the target document using a processor, said likelihood being based on the interest model and the one or more keywords present in the new document; and
projecting an estimated future information spread based on the determined likelihood for each user.

### Claim 10 (depends on 9)

The method of claim 9, weighting a keyword comprises calculating a weight based on a proportion of documents including the keyword that were re-shared by the user.

### Claim 11 (depends on 10)

The method of claim 10, wherein weighting a keyword comprises increasing the weight of a re-shared keyword in inverse proportion to a level of activity for the user.

### Claim 12 (depends on 10)

The method of claim 10, wherein weighting a keyword comprises calculating a weight according to:
q
j
(
u
)
=
∑
m


r
(
m
)

f
j

(
m
)
∑
m


a

(
t
m
)

f
j

(
m
)
+
C
,
where r(m) is 1 if a post m was re-shared by the user u and 0 if the post m is ignored, where fj(m) is 1 if the keyword j appeared in a post m and is 0 otherwise, and where C is a small constant.

### Claim 13 (depends on 9)

The method of claim 9, wherein determining a likelihood comprises calculating a probability that the user will re-share the target document as:
P
(
u
,
m
)
=
a

(
t
m
)

(
1
-
∏
j


(
1
-
q
j
(
u
)

f
j

(
m
)
)
)
,
where fj(m) is 1 if the keyword j appeared in a historical document m and is 0 otherwise and where qj(u) is a weighted value for the keyword j for the user u.

### Claim 14 (depends on 9)

The method of claim 9, further comprising determining a time window in which edge user is most likely to re-share the target document.

### Claim 15 (depends on 9)

The method of claim 9, further comprising repeating said steps of determining a likelihood for each edge user and projecting an estimated future information spread, where the edge users in each iteration are based on the estimated future information spread of the previous iteration.

### Claim 16 (depends on 9)

A non-transitory computer readable storage medium comprising a computer readable program for estimating a probability of re-sharing information, wherein the computer readable program when executed on a computer causes the computer to perform the steps of claim 9.

### Claim 17

A system for estimating a probability of re-sharing information, comprising:
a historical data table comprising a social graph map and a set of documents that each of a plurality of users have been exposed to;
a keyword extraction module configured to extract keywords from a set of documents addressed to each user;
an interest model generator configured to weight the extracted keywords according a metric for the user's interest in the keywords' respective source documents and to create an interest model from the weighted keywords; and
a probability estimator comprising a processor configured to determine a likelihood that the user will re-share a new document, said likelihood being based on the interest model and one or more keywords present in the new document.

### Claim 18 (depends on 17)

The system of claim 17, wherein the interest model generator is further configured to calculate a weight for a keyword based on a proportion of documents including the keyword that were re-shared by the user.

### Claim 19 (depends on 18)

The system of claim 18, wherein the interest model generator is further configured to increase the weight of a re-shared keyword in inverse proportion to a level of activity for the user.

### Claim 20 (depends on 17)

The system of claim 17, wherein the probability estimator is further configured to calculate a probability that the user will re-share the new document as:
P
(
u
,
m
)
=
a

(
t
m
)

(
1
-
∏
j


(
1
-
q
j
(
u
)

f
j

(
m
)
)
)
,
where fj(m) is 1 if the keyword j appeared in a post m and is 0 otherwise and where qj(u) is a weighted value for the keyword j for the user u.