PDF Archive

Easily share your PDF documents with your contacts, on the Web and Social Networks.

Share a file Manage my documents Convert Recover Search Help Contact



US7693827 .pdf



Original filename: US7693827.pdf
Title: US7693827.pdf

This PDF 1.3 document has been generated by / ImageMagick 6.6.0-1 2010-03-04 Q8 http://www.imagemagick.org, and has been sent on pdf-archive.com on 21/05/2014 at 02:15, from IP address 178.162.x.x. The current document download page has been viewed 515 times.
File size: 2.3 MB (29 pages).
Privacy: public file




Download original PDF file









Document preview


US007693827B2

(12) United States Patent

(10) Patent N0.:
(45) Date of Patent:

Zamir et a1.

(54)

PERSONALIZATION OF PLACED CONTENT
ORDERING IN SEARCH RESULTS

(56)

(73) Assignee: Google Inc., Mountain View, CA (US)
Notice:

Subject to any disclaimer, the term of this
patent is extended or adjusted under 35

U.S.C. 154(b) by 492 days.

Apr. 6, 2010

References Cited
U.S. PATENT DOCUMENTS

(75) Inventors: Oren Eli Zamir, Nyack, NY (US);
Jeffrey L. Korn, New York, NY (US);
Andrew B. Fikes, Los Altos, CA (US);
Stephen R. Lawrence, Mountain View,
CA (US)

US 7,693,827 B2

5,724,567 A

3/1998

Rose et a1. ................ .. 395/602

(Continued)
FOREIGN PATENT DOCUMENTS
EP

1050830 A2

11/2000

(Continued)
OTHER PUBLICATIONS
Brin, S., et al., “The Anatomy of a Large-Scale Hypertextual Web
Search Engine,” Computer Networks and ISDN Systems, vol. 30, No.
1-7, Apr. 1998, pp. 107-117.

(Continued)
(21) Appl. N0.: 10/s90,s54
(22)

Filed:

Primary Examinerilohn E Breene
Assistant ExamineriThu-Nguyet Le
(74) Attorney, Agent, or FirmiMorgan, Lewis & Bockius

Jul. 13, 2004

(65)

Prior Publication Data

US 2005/0240580 A1

Oct. 27, 2005

Related US. Application Data

(63) Continuation-in-part of application No. 10/676,711,

LLP

(57)

ABSTRACT

A system and method for using a user pro?le to order placed
content in search results returned by a search engine. The user
pro?le is based on search queries submitted by a user, the

user’s speci?c interaction with the documents identi?ed by

?led on Sep. 30, 2003.

the search engine and personal information provided by the

(51)

Int. Cl.

G06F 17/30

(2006.01)

(52)

US. Cl. ...................... ..

(58)

Field of Classi?cation Search ..................... .. None

707/3; 707/4; 707/5; 707/7

user. Placed content is ranked by a score based at least in part
on a similarity of a particular placed content to the user’s
pro?le. User pro?les can be created and/or stored on the client
side or server side of a client-server network environment.

See application ?le for complete search history.

29 Claims, 13 Drawing Sheets

Previous Search
Queries Submitted

by the User

10.3
URLs identi?ed by
the Previous Search
Queries

lQi
Anchor Text of the
Identi?ed URLs

211.
Term-based Pro?le

221
General Information
about the identi?ed
Documents

Zii

Category-based
P rofile
User's Activities on

the Identi?ed
Documents

m
211
Sampled Content
from the identi?ed
Documents

Category Information
about the Identi?ed
Documents

A5
User's Personal
information

2.11

User's Browsing
Patterns

Link-based Pro?le

US 7,693,827 B2
Page 2
US. PATENT DOCUMENTS
5,754,939 A

5/1998

6,012,051 A *

1/2000 Sammon et al.

6,182,068 B1*
6,285,999 B1

1/2001
9/2001

6,327,590 B1
6,385,619 B1

HerZ et al. ................. .. 455/4.2
Culliss ........ ..
Page ........... ..

12/2001 Chidlovskiiet al. ..
5/2002 Eichstaedt et al.

6,421,675 B1

7/2002

Ryan et al. ...... ..

6,535,888 B1*

3/2003 Vijayan et al.

6,606,619 B2 *
6,868,525 B1*
6,892,198 B2

8/2003
3/2005
5/2005

Ortega et al.
SZabo ......... ..
Perisic et al. ..

706/52
707/5
707/5

707/5
. 707/1
. 707/100

707/104.1
..... .. 707/2
715/738
. 707/5

Illinois, Nov. 1999, pp. 391-398.
International Search Report for International Application No. PCT/
US2005/025081, mailed Dec. 2, 2005.
Joachims, T., et al. “Accurately Interpreting Clickthrough Data as
Implicit Feedback,” Proceedings of the 28th Annual Int’l ACM SIGIR
Conference on Research and Development in Information Retrieval,
Aug. 15, 2005, pp. 154-161.

6,895,406 B2

5/2005 Fables et al.

6,912,505 B2

6/2005 Linden et al. ..

7,031,961 B2

4/2006 Pitkow et al.

707/4

Juan, Y-F, et al., “An Analysis of Search Engine Switching Behavior

7,240,049 B2

7/2007

707/3

707/1

Using Click Streams,” Internet and Network Economics Lecture
Notes in Computer Science, vol. 3828, 2005, pp. 806-815.

707/3
707/10
715/517

Ramachandran, P, “Discovering User Preferences by Using Time
Entries in Click-Through Data to Improve Search Engine Results,”

2002/0073065 A1

Kapur ......... ..

6/2002 Inaba et al. .

2002/0123988 A1
9/2002 Dean et al.
2002/0198882 A1
12/2002 Linden et al.
2003/0149937 A1* 8/2003 McElfresh et al.
2004/0044571 A1

3/2004

707/102

Haveliwala, T.H., “Topic-Sensitive PageRank,” Proc. ofthe 1 1th Int’!
World Wide Web Conf., Honolulu, Hawaii, May 2002.
Jeh, Glen, et al., “Scaling Personalized Web Search,” Stanford Univ.
Technical Report, 2002.
Pretschner, A., et al., “Ontology Based Personalized Search,” Proc.
11th IEEE Int’! Conf on Tools with Arti?cial Intelligence, Chicago,

705/14

Bronnimann et al. ....... .. 705/14

707/103 R

Discovery Science Lecture Notes in Computer Science, Arti?cial
Intelligence, vol. 3735, 2005, pp. 383-385.
Zhao, M., et al., “Adapting Document Ranking to Users’ Preferences

2004/0267806 A1 *

12/2004 Lester

2005/0071741 A1
2005/0144193 A1
2005/0240580 A1

3/2005 Acharya et al.
6/2005 HenZinger
10/2005 Zamir et al.

715/500
707/103
707/4

Using Click-Through Data,” Information Retrieval Technology Lec

2007/0088692 A1
2007/0088693 A1

4/2007 Dean et al. ..
4/2007 Lawrence

707/5
. 707/5

“Yahoo! Search Builder-Design Search Box,” http://webarchine.

2007/0094254 A1

4/2007 Cutts et al.

2007/0094255 A1

4/2007 Acharya et al. .............. .. 707/5

707/5

FOREIGN PATENT DOCUMENTS
EP

W0
W0
W0
W0
W0

1 107 128 A1

W0 03/107127
WO 2005/001719
WO 2005/033979
WO 2005/055015
WO 2006/014562

A2
A1
A2
A1

6/2001

12/2003
1/2005
4/2005
6/2005
2/2006

OTHER PUBLICATIONS
Cho, J ., et al., “Ef?cient Crawling Through URL Ordering,” Com
puter Networks and ISDN Systems, vol. 30, No. 1-7, Apr. 1998, pp.
161-171.

ture Notes in Computer Science, vol. 4182, 2006, pp. 26-42.

org/web/2006813082935/http://builder.search.yahoo.com/m/
promo, Aug. 13, 2006, 1 page.
“Guide to Custom Search Engines (CSEs),” http://web.archive.org/
web/2006102723 5927/http://www.customsearchguide.com, Oct.
27, 2006, 1 page.
Ding, J ., et al., “Computing Geographical Scopes ofWeb Resources,”
Proceedings of the 25th VLDB Conf., Cairo, Egypt 2000.
HenZinger, M., “Web Information Retrievalian Algorithmic Per
spective,” Lecture Notes in Computer Science, Proceedings of the 8th

Annual European Symposium, Saarbruken, Germany, Sep. 2000, pp.
1-8.

International Search Report for International Application No. PCT/
US07/065710, mailed Nov. 12, 2007.

* cited by examiner

US. Patent

Apr. 6, 2010

Sheet 1 0f 13

US 7,693,827 B2

Client-Server Network Environment 100 \~

Contet Server
106

Client 1
102

Search Engine
104

Client 2
102

Fig. 1

User ro?le
Server
108

Client N
102

US. Patent

Apr. 6, 2010

US 7,693,827 B2

Sheet 2 0f 13

A
Previous Search

Queries Submitted

by the User

_20_3
URLs Identi?ed by
the Previous Search
Queries

20_5
Anchor Text of the
Identified URLs

31
Term-based Pro?le

201
General Information
about the identi?ed
Documents

Q

30
User Pro?le )

Category-based
Profile

m
User’s Activities on

the Identi?ed

Documents
1

m
Sampled Content
from the Identi?ed
Documents

_2_1_3
Category Information
about the Identi?ed
Documents

&
User's Personal
Information

Fig. 2
User's Browsing

Patterns

235



\ Link-beg Pro?le )



US. Patent

Apr. 6, 2010

Sheet 3 0f 13

US 7,693,827 B2

Term-based Pro?le Table 300
320

340

H

/

A

\

USER_ID (TERM_1,WEIGHT_1) (TERM_2,WEIGHT_2) .

.

310

Fig. 3

Link-based Pro?le Table 500

530
LINK ID

510

540
WEIGHT

,

\

5104f

Y

5204f USER_1

f USER 2

520-2

/

LINK_|D WEIGHT

_

520-N/ USER—NI '\
LINK_|D WEIGHT

510-Nf

. (TERM_N,WEIGHT_N)

US. Patent

Apr. 6, 2010

Sheet 4 0f 13

Categogi Mag 400

US 7,693,827 B2

Movie

_

(1.1.1)

Lyrics
(1.1.2.1)

Art

Music

News

(1.1)

(1.1.2)

(1.1.2.2)

Literature

(1.1.2.3)

Reviews

(1.1.3)
_

Company

(1.2.1)
_ Business

Investing

(1.2)

(1.2.2)
Job

(1.2.3)
_

Aerobics

Fitness

(13-1-1)

(1.3.1)

Yoga
(1.3.1.2)

Health

Medicine

(1.3)

(1.3.2)
_

Diets

--

Root

__

1.3.3.1

__ Nutrition _

(1-3'3)
_

(

. )

_Shopping
(1.3.3.2)

Journal

(1.4.1)

__

Music

(1.4.2.1)
News

Radio

(1.4)

(1.4.2)

__

News

(1.4.2.2)
l_ Talk Shows

__

TV

(1.4.2.3)

(1.4.3)


Biology

1.

.

( 51)

.

.

Air Quality

(1.5.2.1)

Science

Environment

EcolOgy

(1-5)

(1.5.2)

(1.5.2.2)

Physics
(153)

Noise

Basketball

(1.6.1)
Sports

Football

(1.5)

(1.6.2)
Soccer

(1.6.3)

Fig. 4A

(1.5.2.3)

US. Patent

Apr. 6, 2010

Sheet 5 0f 13

US 7,693,827 B2

Categorv-based Pro?le Table 450

CATEGORY_ID WEIGHT

342-1

f USER_1

3422f

USER 2

1
./

0.2
0.9

1.5.2

0.1

/’NCATEGORY_|D WEIGHT

'

342_nf USER—N

1.1
1.1.3

'\

12
'
1.3.1.2

03
'
0.3

1.6.3

0.7
.
.
.

CATEGORY_ID WEIGHT

Fig. 4B

1.1.2
1.3

045
0.6

1.42.2

0.05

US. Patent

Apr. 6, 2010

l

Sheet 6 0f 13

Begin

US 7,693,827 B2

)

ll

Remove items like comments, JavaScript
and style sheet, etc., from a document

61
f0

ll

Select the ?rst N words from each paragraph f 620
of length >= MinParagraphLength

V

(Optional) Add the document title to the sampled content, f 630
ifthe length of sampled content < a prede?ned threshold

V

(Optional) Add the non-inline HREF links to the sampled content, f 640
if the length of sampled content < the prede?ned threshold

V

(Optional) Add the ALT tags to the sampled content, if f 650
the length of sampled content < the prede?ned threshold

V

(Optional) Add the meta tags to the sampled content, if _ f 670

the length of sampled content < the prede?ned threshold

End

Fig. 6

US. Patent

Apr. 6, 2010

Sheet 7 0f 13

US 7,693,827 B2

List of prede?ned important terms r712
List of prede?ned unimportant terms f 714

(optional)

Context Analysis

Training documents

Training Phase 701

‘V

I710

Receive a set of training data, including a

list of predefined important terms,
(optionally) a list of prede?ned unimportant
terms, and a set of training documents



f 720

Analyze training documents to identify a
plurality of context patterns (pre?x/post?x/
combination), each pattern having an
associated weight

Operational Phase 703 <

v

f

Apply the identified context patterns to a set
of documents identified by the user

‘V

f 740

identify another set of important terms that
characterize the user's speci?c interest

f716


Related documents


PDF Document reliable trademark attorneys malaysia
PDF Document patent registration malaysia
PDF Document amit kleinmann cv for acm
PDF Document summary statement professional engineer cdr writing 1
PDF Document engineers migration guide for australia reviewmycdr
PDF Document how can i patent my1036


Related keywords