(12) United States Patent (10) Patent No.: US … · (73) Assignee: Lockheed Martin Corporation ... Semantic Web, held at ECAI-02, Lyons, France ... “Semi-Automatic Ontology Eng

USOO7987088B2

(12) United States Patent (10) Patent No.: US 7.987,088 B2 Moitra et al. (45) Date of Patent: Jul. 26, 2011

(54) SYSTEMAND METHOD FOR AUTOMATING 3.ES A. : &3. Santa et al. ................. 29,R ta et al. .........

SNNESSEE 2004/0243645 A1 12/2004 Broder et al. .. 707/2OO 2005/0210008 A1 9/2005 Tran et al. ......................... 707/3 2006/0020466 A1 1/2006 Cousineau et al. 704/257

(75) Inventors: Abha Moitra, Scotia, NY (US); 2006/0031288 A1 2/2006 Ter Horst et al. .. TO9.204 Andrew Walter Crapo, Scotia, NY 2006.0053151 A1 3/2006 Gardner et al. .... TO7,102 (US); Michael Allen Bodkin, Orlando 2006/0053175 A1 3/2006 Gardner et al. ............... 707,203 FL (US) s s 2006/0059028 A1 3/2006 Eder ................................. 705/7

OTHER PUBLICATIONS (73) Assignee: Lockheed Martin Corporation,

Bethesda, MD (US) ISI Ontosaurus from http://www.isi.edu/isd/ontosaurus.html (Oct. s 11, 2006); 4 pages.

(*) Notice: Subject to any disclaimer, the term of this Air Tasking Order. From Teknowledge at http://reliant, teknowledge. atent is extended or adjusted under 35 com/DAML/ATO Ontology.owl (Oct. 11, 2006); 10 pages.

p S.C. 154(b) bV 1289 da Berners-Lee, T., Hendler, J., Lassila, O.: “The Semantic Web:” Sci M YW- (b) by yS. entific American, (May 17, 2001). Available online at http://www.

Scientificamerican.com/article.cfm? articleID=00048144-1OD2 (21) Appl. No.: 11/491,627 1C70-84A9809EC588EF21&catID=2 (Oct. 11, 2006): 10 pages.

Brickley, D., Guha, R.V.: “RDF Vocabulary Description Language (22) Filed: Jul. 24, 2006 1.0: RDF Schema' (2004). Available online at http://www.w3.org/

O O TR/rdf-schema (Oct. 11, 2006); 24 pages. (65) Prior Publication Data British National Corpus. http://www.natcorp.ox.ac.uk/ (Oct. 11,

2006); 1 page. US 2008/OO217OO A1 Jan. 24, 2008 Buitelaar, P., Sintek, M.: “OntoLT version 1.0: Middleware for Ontol

(51) Int. Cl ogy Extraction from Text” 3rd International Semantic Web Confer ence, (2004); 4 pages.

G06F2L/00 (2006.01) GOL 5/00 (2006.01) (Continued)

(52) U.S. Cl. .............. 704/10; 704/9; 707/699; 707/705; r 707/711 Primary Examiner — Justin Rider

(58) Field of Classification Search ..................... 70410 (74) Attorney, Agent, or Firm — Withrow & Terranova, See application file for complete search history. PLLC

(56) References Cited (57) ABSTRACT

U.S. PATENT DOCUMENTS

6,076,088 A * 6/2000 Paik et al. ......................... 707/5 6,101,492 A * 8/2000 Jacquemin et al. ............... 707/3 6,263,335 B1* 7/2001 Paik et al. ......................... 707/5 6,363,378 B1* 3/2002 Conklin et al. ................... 707/5 6,389,409 B1* 5/2002 Horovitz et al. .................. 707/2 6,405,190 B1* 6/2002 Conklin ............................ 707/3 6,415,283 B1* 7/2002 Conklin ............................ 707/3 6,513,027 B1 * 1/2003 Powers et al. ... TO6/47

ONTOLOGY CREATION SOFTWARE

ONTOLOGY 28a REVIEW

16

r SOFT. - 30

Systems and methods for the Substantially automatic creation of ontologies from unstructured documents identify phrases and core noun phrases from the respective documents. Links can be extracted from the documents. Concepts can be iden tified from the documents. Ontologies can be automatically created for the documents. The processing is domain inde pendent.

25 Claims, 39 Drawing Sheets

US 7.987,088 B2 Page 2

OTHER PUBLICATIONS

Faure, D., Nedellec, C. A.: “Corpus-based Conceptual Clustering Method for Verb Frames and Ontology Acquisition:” ICLREC Work shop Granada, Spain (1998); 29 pages. “Jena 2–A Semantic Web Framework.” Available online at http:// www.hpl.hp.com/semweb?ena2.htm (Oct. 11, 2006); 1 page. Horridge, M.: “OWLViz—A Visualization Plugin for the Protégé OWL Plugin.” Available online at http://www.co-ode.org/down loads/owlviz co-ode-index.php (Oct. 11, 2006); 3 pages. Kaon 2005 Tool Suite, see home page at http://kaon...semanticweb. org/ (Oct. 11, 2006); 2 pages. Lenat, D.B.: “CYC: A Large-scale Investment in Knowledge Infra structure. Communications of the ACM:” vol. 38, No. 11. (Nov. 1995) p. 32-38. Maedche, A., Siaab, S.: "Semi-automatic Engineering of Ontologies from Text.” In Proceedings of the 12th International Conference on Software Engineering and Knowledge Engi-neering, (2000); Avail able online at http://www.aifb.uni-karlsruhe.de/WBS/Publ/2000/ seke amasst 2000.pdf (Oct. 11, 2006); 8 pages. McGuinness, D.L., van Harmelen, F.: “OWL Web Ontology Lan guage Overview.” Available online at http://www.w3.org/TRowl features (Oct. 11, 2006); 17 pages. Miller, G.A.: “WordNet: A Lexical Database for English:” Commu nications of the ACM, 38(11) (Nov. 1995) p.39-41; Available online at http://delivery.acm.org/10.1145/220000/219748/p39-miller.pdf? key1=219748&key2=2592383511&col=GUIDE&dl=GUIDE &CFID=15151515&CFTOKEN=6184618 (Oct. 11, 2006). Miller, E., Swick, R., Brickley, D.: “Resource Description Frame work (RDF):” Available online at http://www.w3.org/RDF (Oct. 11, 2006); 7 pages. Miller, E., Hendler, J. Web Ontology Language (OWL); Tools, Projects, and Applications; Available online at http://www.w3.org/ 2004/OWL/Hiprojects (Oct. 11, 2006): 3 pages. Missikoff, M. Navigli, R., Velardi, P: “Integrated Approach to Web Ontology Learning and Engineering.” IEEE-Computer Nov (Nov. 2002) p. 60-63. Monte, T. et al.: “T.: Doodle-Owl: OWL-based Semi-automatic Ontology Development Environment.” Prepared for 3rd Interna tional Semantic Web Conference (2004) p. 33-36. 19. Nelson, S.J. et al., Abstract: “Identifying Concepts in Medical Knowledge.” In Proc. of 8th World Congress on Medical Informatics (MEDINFO), (1995) p. 35-36. Noy, N. F.: “Prompt” Available online at http://protege. Stanford.edu/ plugins/prompt/prompt.html (Oct. 11, 2006); 2 pages. Noy, N.F. McGuinness, D.L.: “Ontology Development 101: A Guide to Creating Your First Ontology.” Stanford Knowledge Systems Laboratory Technical Report KSL-01-05 (2001); Available online at http://protege. Stanford.edu/publications ontology development/ ontology 101-noy-mcguinness.html(10/11/2006): 21 pages. Pease, A.: “Suggested Upper Merged Ontology (SUMO).” Available online at http://www.ontologyportal.org (Oct. 11, 2006), 2 pages. Protégé.: “Welcome to Protégé:” Available online at http://protege. Stanford.edu/ (Oct. 11, 2006), 2 pages. Sleeman. D. et al.: "Enabling Services for Distrib-uted Environ ments, Ontology Extraction and Knowledge Base Characterization.” Proceed-ings of Workshop on Knowledge Transformations for the Semantic Web, held at ECAI-02, Lyons, France, Jul. 2002: 8 pages. Stanford Medical Informatics. Protégé OWL Plugin: Ontology Edi tor for the Semantic Web. Available online at http://protege. Stanford. edu/pluginslowl? (Oct. 11, 2006): 1 page.

Stuart, I.: “XML Namespaces Explained:” (Nov. 27, 2002), Available online at http://www.sitepoint.com/article/xml-namespaces-ex plained (Oct. 11, 2006); 9 pages. Tomokiyo, T., Hurst, M. A.: 'A Language Model Approach to Keyphrase Extraction.” In ACL-2003 Workshop on Multiword Expressions: Analysis, Acquisition and Treatment; 8 pages. Weischedel, R., et al. "Coping with Ambi-guity and Unknown Words through Probabilistic Models;” Computational Linguistics 19(2) (1993) 359-382. Welty, C. Guarino, N.: “Support for Ontological Analysis of Taxo nomic Relationships.” Data and Knowledge Eng. 39(1) (2001) 51-74. Yang, H., et al.: “Extracting Ontologies from Legacy Systems for Under-standing and Re-engineering.” In IEEE Int’l Conf on Com puter Softwre & Applicatins, 1999, 6 pages. Schorr, H.; Final Report: “Ontologies, KnowledgeBases and Knowl edge Management.” USC Information Sciences Institute; (Apr. 9. 1997): 76 pages. Cycorp, Inc.: “What is Cyc?” (prior to Jul. 24, 2005); Available online at http://wwww.cyc.com/cyc/technology/whatiscyc (Feb. 16, 2006); 2 pages. “Semantic Knowledge Source Integration.” Cycorp, Inc. (2004); 1 page. “Cyc Enhancement of Informaiton Extraction.” Cycorp, Inc. (2004); 4 pages. “Aerotext(TM); Technology for Extracting Intellegence from Text:” Lockheed Martin Corp. (2005); 2 pages. “AeroText Products Executive Summary Information.” Lockheed Martin Corp. (May 2003) 21 pages. Smith, B.: “The Ontology of Documents and the Technology of Identification.” CERIAS (Nov. 2, 2005); 2 pages. “Semi-Automatic Ontology Eng and Ontology Supported Document Indexing in a Multilinguar Environmnt at http://www.diplomica. com/db/diplomabeitenó905.html (11/15/05) 6 pages. Birgitta Konig-Ries, M.K.: "An Ontology-Based Document-Space as an Adaptable User Interface for Mobile Information Systems;” 5 pages (prior to Jul. 24, 2005). Motta, E. et al. “Ontology-Driven Document Enrichment: Prin ciples, Tools and Applicaitons.” To appear in the Int'l J. of Human Comp. Studies, 41 pp. (prior to Jul. 24, 2005). Lockheed Martin, Inc., “AEROTEXT TM”, Technology for Extracting Intelligence from Text, Sales Pamphlet, (C) 2005, 2 pages. Lockheed Martin, Inc., “AeroText TM Products Executive Summary Information', (C) 2001-2003, 21 pages. Cerias, Purdue University, “The Ontology of Documents and the Technology of Identification'. Seminar, Nov. 2, 2005, Speaker-Barry Smith, Pamphlet, 2 pages. Absolventen Kongress, Nov. 23/24, 2005, “Katalog” for Semi-auto matic ontology ontology Supported document indexing in a multilin gual environment, Diplomica GmbH, www.diploma.de, 6 pages. Michael Klein Birgitta Konig-Ries, “An Ontology-Based Document Space as an Adaptable User Interface for Mobile INformation Sys tems”. University of Karlsruhe, Germany, http://ww.ipd.uni karlsruhe.de/DIANE, 5 pages, published on or before Jun. 30, 2005. Enrico Motta, Simon Buckingham Shum and John Domingue, “Ontology-Driven Document Enrichment: Principles, Tools and Applications'. Knowledge Media Institute. The Open University, Milton Keynes, UK, {e.motta, S.buckingham.shum, j.b. dominque}(a) open.ac.uk, 41 pages, published on or before Jun. 30. 2005.

* cited by examiner

U.S. Patent Jul. 26, 2011 Sheet 1 of 39 US 7.987,088 B2

ONTOLOGY CREATION SOFTWARE

ONTOLOGY REVIEW


FIG. 2A 1. 102 LOGIN

SELECT TAB: CREATE 104 SELECT TAB SELECT AB: NEW SESSION OR SELECT ADD CONCEPTS REVIEWICOMPARE AN EXISTING SESSION TOONTOLOGY ONTOLOGIES

SELECT TAB: EXTRACT O6 CONCEPTS FROM TEXT DOCUMENTS

NAVIGATE AND SELECT INPUTTEXT DIRECTORY

110 SET CUTOFF AND PHRASE LENGTH

NAVIGATE AND SELECT CONCEPTS AND ONTOLOGY DIRECTORIES

NAVIGATE AND SELECT ONTOLOGY DIRECTORY AND FILE

144

VIEW ONTOLOGY IN PROTEGE

146

108

GENERATE INTIAL ONTOLOGY FROM SINGLE-SENSE PHRASES IN FILE Combined File.txt.

112 PhrFreq.csv COMPARE WITH RISE ONTOLOGY SN ANOTHER ONTOLOGY

FILE WITHNAME N PROTEGE USING File.txt PhrFreq.csv) OF THE FORM Ont.OW PROMPTTAB

ACTUAL FILE 114 NAME REFLECTS - CURRENT TIME SET CUTOFF FOR

CORE NOUN PHRASES

EXTRACT CORE NOUN PHRASES (INTO FILE

CombinedFile.txt. CNPCSv)

SET SUBJECT AND OBJECT CUTOFF

SETSUBJECT-PREDICATE AND PREDICATE OBJECT GAP

148

SAVE SESSION

LOGOFF

END

116 ADD CORE NOUN PHRASE CONCEPTS TOONTOLOGY GENERATED IN PREVIOUS STEP

ADD LINK CONCEPTS TOONTOLOGY GENERATED IN PREVIOUS STEP

122 EXTRACT LINKS (INTO FILE CombinedFile.txt. Linktext)

US 7.987,088 B2 Sheet 3 of 39 Jul. 26, 2011

eZOZ

U.S. Patent

US 7.987,088 B2

- - - - - - - - - - - - - - - -

U.S. Patent


FIG. 2G PROCESS LST POSSOntSubCaSS

FROM "YES" BRANCH OF 604

ALL ENTRIES IN LIST POSSOntSubClass PROCESSED

re

SELECT NEXTENTRY OF THE FORM "P1

ASSOCATED WITH P2"

ASSOCATED WITH P2" EXISTS IN POSSOntSubClass AND P3 SAN ANTONYM

MOVE"P1 ASSOCATED WITH P2" TO LIST

PossConceptSubClass

CREATE "P1 P2"AND "P3 P2"AS SUBCLASSES OF P2 IN ONTOLOGY

ASSOCATED WITH P2" EXISTS IN LIST

PossConceptSubClass

DROP "P1ASSOCATED WITH P2" FROM LIST POSSConceptSubClass

F“P3 ASSOCATED WITH P2" EXISTS IN LIST

PossConceptSubClass

DROP "P3ASSOCATED WITH P2" FROM LIST PossConceptSubClass


FIG. 2H PROCESS LSTPossConceptSubClass

1. 700-2 FROM 624

ALL ENTRIES IN LIST

PossConceptSubClass PROCESSED

SELECT NEXT ENTRY OF THE FORM "P1

ASSOCATED WITH P2"

F“P3 ASSOCATED WITH P2"

EXISTS IN PossConceptSubClass AND P3 SAN ANTONYM

ADD"P1 P2"TO LIST PossConcept

DOES P2 EXIST IN ADD P2ASA SUBCLASS

ConceptsToReview OF ConceptsToReview IN ONTOLOGY? NONTOLOGY

CREATE "P1 P2"AND"P3 P2"AS SUBCLASSES OF P2 IN ONTOLOGY


FIG. 2 PROCESS LSTPossConcept

re FROM 628

ALL ENTRIES IN LIST PossConcept PROCESSED

SELECT NEXT ENTRY Phr

DOES Ph EXIST IN BRITISH

NATIONALCORPUS (BNC) WORD LIST?

ADD Phr ASA SUBCLASS OF ConceptsToReview

IN ONTOLOGY IS NORMALIZED

FREQUENCY OF Phr 2-5 TIMES ITS BNC FREQUENCY?


FIG. 2J ADD LINK CONCEPTS TO ONTOLOGY

1. 800

ALL ENTRIES INFILE

CombinedFile.txt. Link.text PROCESSED

GO SELECT NEXT ENTRY 806

<ConfSubj, Pred. Obj, Phr, Catid,SentFrag->

808

DOES Pred EXIST IN ONTOLOGY

YES

DOES Pred HAVEASINGLE

SENSE WHICH IS VERB IN WORDNET

CREATE AN INSTANCE OF LinkSTOReview IN ONTOLOGY WITH 7 ATTRIBUTES

CORRESPONDING TO <Conf. Subj, Pred.Obj, PhrCatid,SentFrag

CREATE WNWordSense FOR EACH WORD IN THE VERB SYNSET IN ONTOLOGY

CREATE ObjectProperly FORVERB SYNSET AND ALL OF ITS PARENTS IN ONTOLOGY:

If Sub ISA CONCEPT IN ONTOLOGY THEN DOMAIN OF THE Object Property IS SETTo Subj;

IF ObjSA CONCEPT IN ONTOLOGY THENRANGE OF ObjectProperty ISSET TO Obj


Ele Edit view Fay crites cols help

Oba - O -(z) (2) as Ossie : Facts e 2- a a 8 Addresse http:ffocal host:8080/OLSijsp/logon.jsp WGo Googlev w search v & 880 blocked & Check v & Autolink v SAutcFi Edoptions 4

Local treet


File Edit View favorites tools Help As

OBack w e) x ? AOsearch Favorites 8 2. a Fr 8. Addresse http:liocaihost:8080/OLSijspliogon.jsp w Go Googlev w G search v 3 880 blocked if Check v. & Autolink v S. Autofit options 40

User Rote:

Locategie


File Edit view Fayorites tools Help

Oak - O -|x|) as osae trfare e 2- a 2 & Addresse http:liocahost:8080/OLSicontrollerinteraction.jspa Googlev 2G Search v 3 sh90blocked ‘y Check v. & Autolink v Frame Selector ------

Create Newnstance Session created 38.06 8:50AM

9 locaterist

U.S. Patent Jul. 26, 2011 Sheet 20 Of 39 US 7.987,088 B2

Fife Edit view Fisyfites locis Help

OBack v Ov A Osearch 5: Favorites & (2v se Evil 8. Addresse http:tioca?hiost:8080/OLSicontrolierinteraction.jspa

G search v é 380 blocked ; Check v: Autolink v is AutoFit doptions & www.

Wedia 08:48 S2O6

Extract Concepts for text octaerts 2A Concepts to Oto 3. ReviewCompare Ontologies

Local internet


ise - Six

Ele Edit view favorites tools Help A5

Ossa Ox) () as Osale yi raone e 2- e a J & Addresse http:llocalhost:8080/OLS controllerinteraction.jspa w s

Googlev G Search v. & 8,80blocked ‘y Check v & Autolink v AutoFill Ecoptions & - E. --- W

G As --

1. Extract Concepts from Text Documents 2. Add Concepts to Ontolo 3. Review/Compare Ontologies s

Etext File Preview e Sest.txt

Please selest text doc directory (currently CSlotsdataio Sfest) Department of Defense

Hickam AFB is located near Holio the state capital of Hawaii; C$loisdata/OLSTest ExtractedConcepts responsible formaintenance of

Go up one directory level. stealth bomber. Planned faintenance is scheduled by the system; while the grosuper is O

Select desired action Setect text file

Do This Step Now Skip to Next Step

Aa2 -------

w

e Dore unknown Zo?e (Mixed)


Fig. 3F w x

File Edit view Fayorites tools help s

Oba - 0x) (2) as osae is fate e a se a 8 Addresse http:lioca host:8080/OSlcontrollerinteraction.jspa w Go

Googlev G Search v a gh80 blocked & Check vs Autolink is AutoFit EOptions & m acces

geBGR P

Wetwa (8 clais Si2O6 1. Extract Concepts from Text Documents 2. Add Concepts to Ortolo 3. ReviewiCompare Ontologies

location flexit Documents EGEE text File Preview EGO Sest.txt

leasa select text doc directory currently CSloisdata.io.Sast) Department of Defense

Hickam AFB is located near Holo the state capital of Hawaii; a CSloisdata/OSTestlextractedConcepts responsible formaintenance of

Go up one directory level. stealth bomber. xxxx Sr. Planned maintenance is scheduled

by the system; While the prosuper is w C-f e 9. -P

Select citesire actics

Do This Step Now Go To Previous Step Skip to Next Step

Selectextile

AE W

e Sore unknown Zone mixed)


x six::::::::::A;

File Edit view Favorites pols Halp A5

O Back 0-x) as A search i Favorites & 2" de E1 3 Addresse httpd localhost:8080/OLSlcorticlerinteraction.jspa wego Googlev

Wed Mar 08 0:01:48 ES 2006 f. Extract Concepts from Text Documents

2. Add Concepts to Ontolo 3. ReviewCompare Ontologies

Text File Preview OSTest.txt

location of text Documents

Pease seect text doc directory (currentiy 'CSfoisdatafo Sest sdatafoLSest) Department of Defense

Hickam AFB is located near Holo e the state capital of Hawaii;

CSloisdata/OLSTest ExtractedConcepts responsible formaintenance of Go up one directory level. Stealth bornber.

Extract Links by the system; while the prosuper is

Subject Cut-off', (0-100);

Max words between subject and verb (links only:

Object Cut-off, (C-100):

Max Words between were ad object (links ority):

Select textfile

Ea Select desired action

Do This Step Now Go To Previous Step

s

ww Yawm M- W

lo?e Unknown Zone (Mixed)


File

Ossa - O -|x|) as Addresse

Edit view favorites tools Help

http:llocaliticist:8080OLSlcontroleriteraction.jspa

Googlev

AOsearch y Favorites 8 2" e v 3.

le

Wed Mar 08 0:0:48 ES 2006

1. Extract Concepts from Text Documents 2

Add Concepts to an Ontology

Go up one directory level,

Go up one directory level.

Note: leave file blank to create new file (single-sense only) Select ontology file

elect desired action...(run in

1. Add Single-Sense Phrase Concepts 2. Add Core Noun Phrase Concepts 3. Add Links (Properties

s E3. ReviewCompare ontologies 5

Concepts File Preview -) CombinedFie.txt.Parfreq.csv

A A B C D -

EE w

ontology directory (currently"CSlotsdata/OSTest/Extracted concepts

A. is responsible for 1.4

4 prosuper

Select concepts file

CombinedFile.txtCNP.csv CombinedFile.txt link.text CombinedFile.txt Phrfreq.csv

gunknown zone (Mixed)


File Edit view Fayottes tools Help

OBack Ox) A Osearch Favorites 8 2 e 8 Addresse http:llocalhost:8080ioLSicontrollerinteraction.jspa Googlev

s: e

Wed Mar 08 0:01:48 ES 2006 -------- Add Concepts to Ontology

1. Extract Concepts from text Documents 3. ReviewiCompare Ontologies

Add Concepts to an Ontology c Concepts file Preview -v- ------ ERC)

Combined file.txt.Ph?freq.csy oncepts directory (currently 'cSlolsdata/OLSTestiextractedConcepts")


ology directory (currently "CSioisdata iOSfestiextracted Concepts")

ecies:es


department flighttine awaii iw

Note: leave file blank to create new ?ite (single-sense only) Select ontology file

Select concepts file 1. Add Single-Sense Phrase Concepts 2, Add Core Noun Phrase Concepts 3. Add Links (Properties

e done Unknown 20ne (Mixed)


Ele Edit yiew Fayofites tools help

3. ReviewCompare Ontologies C.

Concepts File Preview C Combined File.txt link text

Add Concepts to an Ontology EG

hickam aft is locat taintenance of planned mainterarce Wing cominander will not prosuper for

ctedConcepts")


Note: Leave file bank to create new file (single-sense only) Select ontology file

Select concepts file 1. Add Single-Sense Phrase Concepts 2. Add Core Noun Phrase Concepts 3. Add Links (Properties


--- nows st --- -- -. s es - S. siteia:a::x: O X

File Edit view faycrites roots Help A7

(OBack ed 6 Osearch Favorites & 27 da Eva & Address & httpifocalhost:8080/OLSlcontrollerinteraction.jspa W Go

Googlev G Search v. & Sh30blocked Check vsk Autolink v Autofil Edoptions & --------------- ---------------- --- ---, s: M

ES BEGR siri s: Ex33:8:

Wed Mar 08 0:01:48 ES 2006 3. Review/Compare Ontologies d

1. Extract Concepts from Text documents 2. Add Concepts to Ontolo -

iew and Compare Ontologies EG.)

intology directory (currently "CSIolsdata/OLSTestlextractedconcepts")


Select ontology file

Select desired action...

1. View Ontology in Protege 2. Compare with Another Ontolog

A. s: -- -

W

e ore Unknown Zone (Mixed)


Ox Ele Edit Project OWL. Code Window Prorapt Tools help

for Project ontí141830226354

Asserted Hierarchy

w O swat; ConceptsToReview O hawaii Ohickama?b. O immediate action_tcto

v () maintenance O planned maintenance O unplanaedrainterarce

Oprosuper O system

O swat:nfoSource OswatLinkstoReview

v () swatwinsynse w O entity

w Oabstract entity D O abstraction

v O physical entity p O causal agency-causal agent cause v Oobject-physical object

D animal. Jiving-living thing b. Olocations v Ounit--whole

v Oartefact-artifact v O instrumentality-instrumentation

w Oconveyance-transport v Ovehicle

b () craft D. Omilitary vehicle

swat:WNWordSense

swat:WNSynset

swat:WNWordSense

K <e protégé

immediate action clo

pianned mai


Ele Edit Project W. Code Window ProTip Tools Help

2 () ré (e. g. as 4 (1 V > t > a is < D. W protégé

OEO.8 c Assetted Model

OWLW2.

for Project ontf141830226354

Asserted Hierarchy

E stealth bomber

-cat salagent cause executive department defense-defense department-department olefense-dod object t-federal departmente-federal offic

g-living thing

-artifact

urrentality-instrumentation conveyance-transport O vehicle D Ocraft v O military vehicle v O military plane-warplane

v Obomber Ostealth bomber


Eile Ecil Project QWL Code Window Prompt iocts -elp

e () leg as 4 (w >e e S < D. W.

For individual 0 wing commander will notify the prosuper (instance of Swat:linkstoReview)

E (t. Eis Annotations

e swat: ConceptsToreview e swat: Infosource (1) e swat: LinkstoReview (5) 69 swat WinSynset é swat: WNWordSense (120)

planned maintenance is schedule prosuper is responsible for unpla unplanned maintenance on theft wing cominander will notify help

e swai; inkstofeview

Name

wing con?tanderwilnotify the prosuper O

fofs:comment

(P & swatsourceContext swathascategory A swathasObject g commander will notify the value typ

AP & CS:38

communication

swathasSubject dP & swathascon?idence 9 x

swat:linkWord O & swat:fromSeatence d ce x

the wing commander w.


Ele Edit project QWL Code Window Prompt tools Help

he e () (a es 4 0 V (2 c () > EJ BJ E -< D. W <Sprotégé screase 0 individuals e.

For Project: 8 ont1141830226354 O for individual 0 wing commander will notify the prosuper (instance of swat:LinkstoReview)

Class Hierarchy

state capital

Asianasietention Anotations { s O

e swat: ConceptstoReview e swat: infoSourco (1) és swat: LinksToReview (5) e swat: WNSynset essagaasenseries

rdfs:continent d 0 dominion-n2 0 entity-n1 0 event-n1 (executive_department-n1 { eye-n- 40 federal department-n 0 federal office-n1 0 fight fine-n1 (geographic are-n1 0 geographic region-1 { geographicalatea-n1 0 geographical region-n1 0government department-ni {} group-n1 0grouping-n1 0 heart-ná () heavier than aircraft-n1 (hiking-n1

{) human action-n1

e state capital e WNWordSense

e


- --- x

File Edit view Favorites tools Help 45

C Back v e) -x) (s AOsearch y Favorites 8 2 e 4. w 8. Addresse http:/tlocalhost:8080/OLSicontrollerinteraction.jspa Wo Go

Googlev v. 6 is 80 blocked 5; Check v. & Autolink v AutoFill Eidoptions & S M

Wed Mar 08 0:0:48 EST2006 1. Extract Concepts from Text Documents (e. 2. Add Concepts to Ontology 3. Review/Compare Ontologies -

location ext Documents D EG O s

Sest.txt Please select text doc directory (currently

CSIolsdata/OLSTest) Department of Defense

s Go up one directory level. Hickal AFB is located near Holo

the state capital of Hawaii; a responsible formaintenance of stealth bomber. Planned faintenance is scheduled by the system; while the prost KTE

E.

Select text

A. s -- w

Ocore Unknown Zona (Mixed)


Eite Edit aroject W. Code Window Prompt Tools Help

e () (e. g. as 4 (W (2 CD » e > a. ES < D. W

RANASING UPFONOLOGES

OCompare your current ontology to a different version of the same ontology.

O Move frames between your current including project and one of the included projects

O Merge two ontologies and add the resulting merged ontology to your current project

OExtract a portion of another ontology and add it to your current project.

W Click hete to begin


Ela Edi Project CWL Code Prompt tools help

e i t (G) g as as W (7 (c) (r. 8 (e) J (a Ba < D. W.

kiANAGENGMUff OfCLOGIES

igow

<protégé

CCompare your currentortology to a different version of the same ontology.

O Move frames between your current including project arid one of the included projects

CMerge two cintologies arid acid the resulting merged ontology to your current project.

OExtract a port:o of another ontology and addilo your current project.

Choose the version to compare with the current project elicioisdatafostestExtracted Conceptston 1141830226354 pp.

Display changes for included frames

select a slot containing a concept ID (optional) if resences -Ouches arease an

W Click hete to


Ox Ele Edit Project OWL Code Wincaw Prompt tools help

2 (R - (a 6 as a { w () (c. It is G 2) is ES 4D

sie Tree view

for Project: 8 orth 418309008 f7

Subclass A a Y X & X owi:thing

w swat Concepts foreview O immediate action_ato

For Class

Ohickarna?h v maintenance

eplannementenanes uplanned maintenanes

O prosuper Qeytara

Oswatinfosource OswattlinkstoReview

v swatWNSynset V (entity

b 8 abstract entity v physical entity

s Ocausal agency-causal agent-cause V object-physical object

b animate hing-living thing v location

b. 8 region

b. O unit...whole OswatwiwordSense

V K A. <Q protégé

For instance,

Changed by: Offerences

Cld Walue


e, , cf. Le 6 as a 4 W () (c. ( > t > ) is as < D. W <protégé

atte view sease

image table A. f. S renamed operation rename explanation

8 No elete <nu> a. planned Taintenariceliss., g Delete anu> fine-- No Gelee crux system No. Delete crud parred maintenance No. Delete wri> unplanted Maintenance). N. Delee Knut

0flight line-n No Delete <rud unplanned maintenarica No Desis <nted ghtine No Detee ant:

prosuper is responsible. Nic beate <ntic 8Ontology(http:ltesearch,ge. Ontology(http:lesearch ge. No Alap Directly-changed frame name and type a?e thes. KXWordNet2thySq 0 WordNet2.1MySq Nic hap Directly-changed frame name and type are the s.

abstract entity-1 4) abstract entity-n1 N ap lsorrorphic frame name and type are the s.l., abstraction-6 (abstractic?-ins No Map lsonophic frame name and type are the S., activity-in {d activi: NG Map sorrorphic frame name aid type are the s...,

(dact-2 (act-n2 o Map isomorphic fare narTeard type are the S. administrative body- {d administrative body-n o kap Isomorphic fafne narie and type are the s.l., achinistrative district-1 (administrative district-n1 ho ap sorrorphic frame name and type are he s. administrative division-int (administrative division-ni la hdap isomorphic frame name and type are he s. administrative ufail-int 0 adttinistrative unit-ni g ap isomorphic frame name ardtype ate the s. aeroplane-n 0 aeroplane-n g isoftcrphic tama name and type are the S., w

Changed by:

Differences Y X. Operation Restriction New Waia


x Ele Ea Project OW. - Code Window Prompt tools Help

e () at a 6 as a W () (c. ( > J J E S < D. W protégé Prompt

are see Tree view

For Project: 8 cat 114183090087

Subclass A, o a Y X R/ X owl:thing

w swat: Concepts oreview

O immediate actionicto Ohickamafb

v (maintenance

For Class: 9 swat:LinksToReview (instance of owl Class) X o X Annotations s

w swatinks foReview For Class:

:NAME

Ohickarna?m is Jocated near honolulu 0 wing commander will notify her prostiper

Changed by:

Differences

: Cid Value procupolic-respongico-four-pla?ingdaoinlargrico. 0 peried-Raiatehaheelescheduled by the syster

w swat:WNSynset v entity

b. O abstract entity w physical entity

Ocausal agency-causal agent-cause v object-physical object

b. O animate thing-living thing v (location

Oregion

b O unit-whole Oswat:wnwordSense

U.S. Patent Sheet 38 of 39 Jul. 26, 2011

Ela Edit fair?t yies tes

Basad on tha chronic cough manageriant guidolina

1) Chronic cough is cough that lasts for at least 3 weeks. Chast radiographs shogid be crge tag before any tastient is prescribed in heary at patianis silh chfe?tic cough (G: side ii-2). Chast gagiographs do not have to be foutinely cbtained tecte beginning treatment. for pie surnsd. PNS (post nasal drip syndrome) in young nonstrokets, in pregaant cricf. of the fore obsefving the fesuit of discontinuation of an ACE- {ACE inhibstos of 4 weeks for patients who gave icpeg cough shotty attoi taginaling to Laika an ACE-i

2} Nota; when the chest x-ray is nona fins, astha. atts GERd (gastice behagiai raisix sisease) are the causes of chtonic coug- in PNES. sinus its may is the causa ssp. to approximataty 39% of the irre was cough is nonproductiva, up to 3prex ten at 8 y 60% of the time whers cough is productive. A negative eccrine adation: Sirius cf scans are not to reiy recoraafsced to swaluata tot sinus is as the tsusa of cough. Four-view sinus radiographs should be of do?ed insaac,

3) White 24 noir escahageal phi incinitoting is line most diagnostica iy satui test of assassing for GERD as the cause of cough, conventional insicas used by

US 7.987,088 B2

gastrocitorologists to assass for escopsagt is ?tay tip in is leadingly Agithai Therefor, urt it futute studits provide totter guidelirios, the test shouid be read as acrona whers conventional indices are within the not mal ranga sfid Ac suspicious ?ofluxing cod coughs appear during the mentering session (Grade il-2}

art-o-A

suspicious refluxinksticed costs appilat during the rnor:oring

PPE Eleous erro 9As agreer assessexesarsal

assassociates straassesses tar $ss to assess fice cxxivarxxsankatavax: as to xarks on Toso saloose gearer it::priors

EE East E

corontorial indicos used ty orms total incoax uses bypassertsevitats to 54 sea w .g., g3ifexisterckgests gastroantafotosis copkayamay to eruasary roast

SEEErreros SEEE ww twww.icw Agwi wakkus riskizakavad cowa Kwai survy who feweg was less 2

Ma Y1

city: raw:Kagyk stokix oved to eaty toxsw:x weacy8x in owly capawerk win ovak was a i-3}

chest racographs shikita ordared before any treatient chast radiographs

arrierracarawasawassasaware

the incast diagnosticasy tissui ast

rarrica

3nt:424 houfaophagkaiph reaso&ngs we maxi &arossay usefva test tweasessing befd as the case occash

ceasanteereeeown sts SEs Essessessssssss SAE

critis isk&hor for 4 was for pickers who cawkeyakcash swot ahr G.75 beginning bc.gifying totaka an aca - texting to take an ace

as NewMSW328i xxxiraxis

e pris of betracterve the resa orgaonaweam ca: acetace inharor) M4 wedishwater two ceased oxyawiyahsa teaway to taxa at azo

Ato

assches is chest x-ray is normal pias Essee testers statestaurosawassessessessee

weaks fox psilants who seweepsa to patients cough cxxh EEA 3 is esses SEE

pnd syndross drig Young ?cistickers in expressarodplai (path&scp stews wisdireckars

Fl "


httsS907S2888

Asserted Hierarchy shstract exity

v C abstraction b attributa v psyctoogical featurs v copition-knowledge-noesis

v () information v avidcnce-grounds

v. Osympirit b informatic-tedness

b. O avert v Ophysical entity v Ocausal agency-causal-agon-causu

b agent v individual-rorai-persor-songbody-soneb w backt-grownup

v () protesscra-professional-person V caragiwer-healthcare-provile?

- Obiedical man-?todkcalpract b ?ter-substance v Ka-gityscal chea

aben?hing-iving thing irr-hoo

v () aftefact-artisa p () creator

Oswat www.criSense

exophagas-e Sophagits

rangedical practsowd

ste-ascrap-shadows ph-diagraa-diagraph

genessed

sophagaarafax-sastroesophagnarefuz-Osgothsparefux

NApritory.giseast-frvpiratory-go-orger-sspitatsrika

doc-dward?, retic-exico-physician

&sharh-ashrattack-blonchist

US 7,987,088 B2 1.

SYSTEMAND METHOD FOR AUTOMATING THE GENERATION OF AN ONTOLOGY FROM UNSTRUCTURED DOCUMENTS

FIELD

The invention pertains to systems and methods for creation of ontologies. More particularly, the invention pertains to domain independent systems and methods that can automati cally process a corpus of unstructured documents into one or more ontologies.

BACKGROUND

An ontology is a model of the important entities and rela tionships in a domain. Ontologies have been used in capturing the semantics of a system. Much of the focus in developing semantics and ontologies has been on handcrafting them. That may be appropriate for high level concepts that are domain-independent and which will be used by virtually all domain specific ontologies, but there is a need for a less time-consuming approach to generating domain ontologies, where extensive knowledge is already captured in unstruc tured texts.

Machine processable domain knowledge has been cap tured in a variety of formats ranging from expert systems to procedural code to UML models. Regardless of the represen tation paradigm, the knowledge has in general been hand coded. On the other hand, large quantities of knowledge have been captured in unstructured texts such as conference pro ceedings, technical papers, books, and more recently web pages and electronic documents. The latter have generally been written by subject matter, or, domain experts while the former have traditionally required a tight collaboration between domain experts and knowledge engineers or pro grammers. It has generally been difficult or impossible to obtain the former from the latter by automated, computer assisted means.

There are a number of ontology extraction tools mentioned in the literature including: ASIUM is a system based on clustering and cooperative

methods that is targeted to extracting taxonomic rela tionships.

ExtraKT extracts ontology information from structured database like Prolog knowledge bases.

Ontol T is a Protégé plugin for ontology extraction from annotated text.

Onto earn learns domain concepts and detects taxo nomic relationships among them to produce a domain concept forest.

TextToOnto is part of KAON tool suite. TextToOnto extracts terms that can potentially be included in the ontology as concepts and it also does rule extraction either on the basis of proximity of two terms or by looking for common patterns like “term like term' to identify hierarchy relationships.

XRA—uses Software reverse engineering approach to extract an initial ontology from given data sources and their application programs.

DOODLE-OWL is a system for on-the-fly ontology con struction. This system relies heavily on user interaction in this construction.

In view of the large quantities of information available in unstructured hard copy texts, web pages and electronic docu ments, there is a continuing need to be able to extract concepts or relationships from existing, unstructured documents. In view of the existing Volumes of unstructured documentation

10

15

25

30

35

40

45

50

55

60

65

2 and the rate of increase, domain independent automated methods and systems are preferable. It would also be prefer able if required human involvement could be minimized.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a block diagram of a system which embodies the invention;

FIGS. 2A-2J are flow diagrams that illustrate processing in accordance with the invention;

FIGS. 3A-3U are exemplary screens that illustrate aspects of the processing of FIGS. 2A-2J; and

FIGS. 4A-4B illustrate link extraction and ontology pro cessing in accordance with the present invention of an exem plary medical document.

DETAILED DESCRIPTION

While embodiments of this invention can take many dif ferent forms, specific embodiments thereof are shown in the drawings and will be described herein in detail with the understanding that the present disclosure is to be considered as an exemplification of the principles of the invention, as well as the best mode of practicing same, and is not intended to limit the invention to the specific embodiment illustrated. A system which embodies the invention can automatically

process, identify and extract semantic domain information from technical unstructured documents. Extracted informa tion is structured as an ontology to provide a semantic repre sentation. The identification and extraction of relevant information is

domain independent. Documents which incorporate concepts from a variety of different domains can be structured into different, domain dependent ontologies using common, domain independent, processing in accordance with the present invention.

In one aspect of the invention; phrases of a predetermined maximum length can be extracted from a document. Core noun phrases can then be extracted. The relevant nouns or noun phrases in the document are automatically extracted, independent of the domain of the document. The extracted nouns and noun phrases represent the key concepts in the document.

Links can be extracted into a file. The information extracted from the document can then be structured into an ontology.

In yet another aspect of the invention, phrase concepts can be added to create a new ontology or added to an existing ontology. Core noun phrase concepts can be added followed by link concepts.

Embodiments of the invention make it possible to extract semantic models (ontologies) from a vast repository of unstructured text, which documents much of human knowl edge. Drawing on this repository as well as Subject matter expertise, the present methodology and system aid in the construction and maintenance of an ontology by the Subject Matter Expert (SME).

In another aspect of the invention, an initial ontology can be generated automatically. Subsequently, the concepts and links that were extracted but not integrated directly into the ontology can be presented to the SME in a drag-and-drop environment so that they can be easily folded into the ontol Ogy.

In a disclosed embodiment of the invention, information is automatically extracted from unstructured documents. An initial ontology can be automatically created using non-do main specific processing. A workflow environment can be

US 7,987,088 B2 3

provided which uses known ontology tools (such as publicly available software, PROTEGE from Stanford Medical Infor matics) for visualization, comparison and maintenance.

FIG. 1 illustrates a system 10 in accordance with the inven tion. Unstructured documents, web pages and the like, all without limitation, generally indicated at 12 are input to and processed by one or more programmable processors 16. Text in various forms can be stored in database(s) 18. Documents 12 can be combined into one file and processed, or, stored in separate files and processed all without limitation.

Processing can be carried out by software recorded on a computer readable medium, having various functions, gener ally indicated at 20. Representative executable modules include Software 22 that acquires input material Such as docu ments, web pages, e-mails and the like all without limitation, in any available form from all available sources. Acquired input material can be stored in database(s) 18 for subsequent processing via database interface Software 24.

Acquired input materials can be initially processed by text extraction software 26 as discussed Subsequently. Records output by extraction software 26 can then be automatically incorporated into an ontology created and/or maintained by domain independent ontology creation Software 28. Struc tured records can be stored in database(s) 18.

The subject matter expert, SME, can interact with the soft ware 28 and structured records in database(s) 18 via graphical user interface software 30 and keyboard/display devices 32a, b. The SME can resolve issues left open by software 28.

It will be understood that the software modules such as 22-30 can be executed by one or more processors which can communicate via one or more computer networks such as the Internet. The particular hardware/software configuration for a selected implementation is not a limitation of the invention.

FIGS. 2A-2J are flow diagrams that illustrate processing in accordance with the invention. FIGS. 3A-3J are exemplary screens that illustrate various aspects of the processing illus trated in FIGS 2A-2J. As those of skill in the art will understand, the processing

illustrated in the flow diagrams of FIGS. 2A-2J could be implemented in the system 10 in various embodiments. For example and, without limitation, processor 16 and intake Software 22 could be located adjacent to a corpus or plurality of documents to be processed. The intake software 22 could communicate with the extraction software 26 via one or more computer networks. Similarly, the ontology creation Software 28 could communicate with the databases 18 also via one or more computer networks. Ontology review software 28a can be used by the SME to review and/or compare ontologies.

FIG. 2A illustrates a domain independent method of pro cessing 100 in accordance with the invention. A user logs in, step 102, via a logon screen, FIGS. 3A and 3B. In a step 104 a user selects a function to either create a new session or select an existing session as illustrated in the screen of FIG.3C. A user elects whether to extract concepts from text documents, add concepts to an ontology or review and compare ontolo gies as illustrated in the screen of FIG. 3D.

In a step 106 the user has elected to extract concepts from text documents. In step 108 the user has selected an input text directory, as illustrated on the screen of FIG. 3E, and has selected a text file a portion of which is displayed on the right half of the screen of FIG. 3E.

Subsequently, in step 110 on the screen of FIG.3E, the user also sets phrase concept cutoff percentages and phrase length parameters. The cutoff percentage specifies a degree of fre quency required to keep a phrase in a sorted list. Also in the screen of FIG.3E the user can elect to extract phrase concepts

10

15

25

30

35

40

45

50

55

60

65

4 from the selected text file, illustrated in step 112. The phrase extraction process is illustrated in more detail in the flow diagram of FIG. 2B.

Subsequent to extracting the phrases from the selected documents, the user can, in step 114, using the screen of FIG. 3F set a percentage cutoff for core noun phrases. The selected percentage cutoff establishes a degree of frequency required to keep the respective core noun phrases in a sorted list. Those that fail to exhibit the selected degree of frequency are dis carded. The process of extracting core noun phrases can then be selected in the screen of FIG. 3F step 116. Additional details of the process of extracting core noun phrases are illustrated in the flow diagram of FIG. 2C. The goal of the process of extracting core noun phrases,

step 116 is to identify textual elements that incorporate any number of adjectives, including Zero, followed by nouns. Processing is implemented by use of files or lists of words, such as adjective words, barrier words and verb words. The barrier words and verbs are words that are not likely to be part of a core noun phrase. The adjective words are words that are usually used as adjectives. The list of barrier words includes those words which force

a separation of phrases. The adjective words list includes those words which primarily appear as adjectives. A word in this list may appear only as an adjective. A list of stop words can be provided. These are words which are to be ignored.

Link extraction is initiated in step 118. As illustrated in the screen of FIG. 3G a user can establish cutoff percentages for Subjects, and objects. Additionally in step 120 gaps, maxi mum words between subject and predicate, can be set by the user as well as maximum words between the predicate and an associated object. The user using the screen of FIG. 3G can initiate the link extraction process, step 122. Details of the link extraction process are illustrated in FIG. 2D. The goal of the link extraction process is to extract rela

tionships and links of the form “subject-predicate-object’. These three components are driven by word lists where the Subject and object words are generated from the input docu ments to be processed. Each extraction has a confidence level associated with it. A publicly available lexical database like WordNet organizes English words into sets of synonyms called synsets and WordNet is utilized in generating these word lists.

Completion of the extract links processing, step 122, results in a file with an unstructured “flat” representation of the important concepts and relationships from the document. In step 116, the processing to extract core noun phrases can be expected to load all information which is extracted from the document into the respective file.

Subsequent steps relate to establishing or expanding an ontology, steps 130-138, to provide a structured representa tion of the information extracted from the document. Subse quently, in steps 140-148 various ontologies can be reviewed or compared.

In step 130 a user can select a tab to add previously extracted concepts to an ontology. The screen of FIG. 3H illustrates both a concepts directory and an ontology directory which can be selected by the user, step 132. Additionally, the user can select as a desired action adding single-sense phrase concepts, core noun phrase, concepts or links. The display of FIG. 3H illustrates contents of the phrase

file, the phrases extracted in step 112. In the upper right of the display of FIG. 3H three columns are presented. These are frequency, normalized frequency and the phrase. Phrases are Sorted by frequency. The phrases illustrated on the display of FIG.3H are those

which were extracted in step 112 from the document illus

US 7,987,088 B2 5

trated on FIG. 3E. In step 134 a new ontology can be created by selecting as the desired action, step one namely, adding single-sense phrase concepts. Additional processing details of step 134 are illustrated in the block diagram of FIG. 2E.

Subsequent to creating a new ontology, step 134 the user can add core noun phrases to that ontology in step 136. The display of FIG. 3I illustrates on the upper right therein the core noun phrases file created in step 116. The core noun phrases have been extracted from the document illustrated in FIG. 3E. The user specifies the core noun phrase file via selecting the appropriate entry in the lower right portion of that display.

The core noun phrases file illustrated in the upper right of FIG.3I once again has three columns, frequency, normalized frequency and the core noun phrase. The phrases are sorted by frequency of occurrence. The link concepts from the file previously created in step

122 can then be added to the ontology, step 138. The display of FIG. 3J illustrates in the upper right a portion of the link text file from step 122. The illustrated portions of the file are the confidence column and “subject sorted by confidence value. Additional details of the processing of step 138 are illustrated in the flow diagram of FIG. 2.J.

In step 140 a user can elect to review and compare ontolo gies, as illustrated in the screen of FIG.3K. When so elected, the user in step 142 can select an ontology directory and file along with specifying whether an ontology is to be viewed or compared with another ontology.

In step 144, as illustrated in the screens of FIGS. 3L, 3M user can view a graphical representation of an ontology, which includes a previously created ontology based on an unstructured document. Additionally, the user can, as illus trated in the screen of FIG. 3N, view extracted links in the ontology. Finally, in the screen of FIG. 30 the user can view captured instance data.

Subsequently, in step 146 the user can compare two ontolo gies. The screens of FIGS. 3P-3U illustrate various aspects of the comparison. In FIG. 3Pan exemplary second file is illus trated which will be compared with the previously structured file. In FIG. 3O a user can select the “compare” tab or button as the next step in the process. In the screen of FIG. 3Ra second ontology can be selected for comparison purposes. Finally, in the screen illustrated in FIG. 3S a comparison of the two ontologies is illustrated via a tree view. In the screen of FIG.3T a comparison of the two ontologies is illustrated with a table view. Finally, in the screen of FIG. 3U a com parison of the links of the two ontologies is illustrated. The flow diagram of FIG. 2B illustrates processing 200 for

extracting phrases, step 112. An exemplary text fragment 202a is provided as well as exemplary phrases there in, 202b, which can be extracted by the processing 200.

In initial step 204 a plurality of input documents 12 can be combined into one file. Then, in step 206 all text between delimiters such as <tag, </tag is deleted. Remaining text is all converted to a lowercase representation.

In step 208 total word count is set to Zero and list L1 is set to empty. In step 210 the available text is split into sentence fragments or chunks based on barrier characters. Barrier char acters include "..!?:” so that extracted phrases will not span any of the barrier characters.

If all sentence fragments have been processed in step 212, then flow goes to step 214 where for each entry in the list L1 the normalized frequency is determined. In step 216 all entries in List L1 having a normalized frequency value below the preset cutoff limit are deleted. Finally in step 218 the list

10

15

25

30

35

40

45

50

55

60

65

6 L1 is sorted based on normalized frequency values to produce the extracted phrases output file and the process returns to step 114, FIG. 2A.

If all sentence fragments have not been processed, step 212 the next sentence fragment is split into words based on non word characters, step 222. In step 224 the total word count is incremented by the number of words in the respective sen tence fragment.

In step 226 all words which incorporate one or more numeric digits are discarded and the list L2 is set to empty. In step 228 all sequences of words up to the user pre-specified maximum phrase length are identified as a phrase and added to the list L2.

In step 232 a determination is made as to whether all entries in list L2 have been processed. If not, in step 234 the next phrase in list L2 is selected.

In step 236 an evaluation is made as to whether the phrase is incorporated in the stop word list. Exemplary stop words include:

Because The A An About Additional stop words could be incorporated as would be

understood by those of skill in the art. If the phrase is in the stop words list, we check if all entries are processed, step 232. Otherwise, in step 240 a determination is made as to whether the phrase is present in list L1. If not, in step 242 the phrase is added to list L1 with a frequency value of 1. If the phrase is already in list L1, in step 244 the frequency value of that phrase in list L1 is incremented by 1. The flow diagram of FIG. 2C illustrates processing 300 for

extracting core noun phrases, step 116. Region 302 of FIG. 2C illustrates examples of identified core noun phrases from the text fragment 202a. Further, as will become apparent from the Subsequent discussion, the core noun phrase processing proceeds without prior knowledge on the part of the method ology 100 as to the relevant concepts of the text being pro cessing. As a result, processing 100 is Substantially domain independent.

In addition to be able to extract concepts from the subject documents, without prior identification thereof, acronyms or abbreviations can also be extracted from the text and treated as concepts. An example acronym is TCTO which appears in the list of extracted core noun phrases 302. Similarly, “pro Super while not in any English dictionary is a concept in the domain of the text fragment 202a of FIG. 2B and can be extracted by the processing 300.

In a step 306 previously extracted phrases in the Com binedFile.txt, step 112 are processed, if not previously stored. Text between tags is dropped and remaining text is converted to lower case, as in step 206.

In step 308 total word count is initialized to zero and list L1 is set to empty or null. Text is split in to sentence fragments based on barrier characters step 310, see step 210.

In step 312 a determination is made as to whether all sentence fragments have been processed. If yes, in step 314, for each entry in the list L1 normalized frequency is deter mined. In step 316 all entries in list L1 with a normalized frequency below the user preset cutoff are deleted. Finally in step 318 the list L1 of remaining entries is sorted based on normalized frequency and the process returns to step 118, FIG. 2A

In the event that the all sentence fragments have not been processed, step 312, in step 322 each sentence fragment is split into words based on non-word characters. In step 324 the

US 7,987,088 B2 7

total word count is incremented by the number of words in the sentence fragment. In step 326 all words with numeric digits are discarded and list L2 is set empty.

In step 328 core noun phrases are identified within the sentence fragment by starting from the left most word in the fragment. A core noun phrase corresponds to the maximal phrase “PHR1 PHR2 in sentence fragment where all words in PHR1 are in the adjective word list 330a. The words in PHR2 are not in the adjective word list330a, the barrier word list 330b or the verb word list 330c. This processing can identify multiple core noun phrases from a single sentence fragment. Where all entries in the list L2 have not been processed,

step 332, the next phrase in the list L2 is selected in step 334, and then it is determined if the phrase is in list L1 in step 336. If not, that phrase is added to the list L1 with a frequency value of 1, step 338. Otherwise, the frequency value of that phrase in list L1 is incremented by 1, step 340. The flow diagram of FIG. 2D illustrates processing 400 of

link extraction step 122. On FIG. 2D extracted links 402 are illustrated for the text fragment 202a.

In an initial step 404 subject and object word lists 404a, 404b are generated from the file CombinedFile.txt.CNP.csv, from step 318, based on previously established subject and object cutoff parameters. These subject and object word lists are domain specific and have been extracted from the docu ment file. The ability to recognize and extract subject words and object words from the respective document(s) contributes to the versatility and flexibility of the link extraction process 122.

Steps 406, 410 correspond to steps 306–310 as previously discussed. If in step 412 all sentence fragments have been processed, the links of list L2 are sorted based on descending confidence step 414.

In step 416 a determination is made as to whether all entries on the list L2 have been processed. If so, the process returns to step 130. If not, the next entry is selected in step 418.

In step 420 a determination is made as to whether “Seq (phrase from Subject thru object) has an odd number of paren thesis. If so, the processing goes to step 416. Otherwise, in step 422, a determination is made as to whether the compo nents “subject”, “predicate” and “object' are already repre sented in list L1 as part of a prior extraction. If so, processing continues to step 416. Otherwise in step 424, the entry is added to CombinedFile.txt.link.text and processing contin ues with step 416.

In the event that all sentence fragments have not been processed step 412 the next sentence fragment is selected and the list L2 is set to empty or null, step 428. Step 430 specifies fragment processing to identify instances of “subject’. “predicate” and “object”.

Subjects and objects are elements that match an entry in the subject file 404a and object file 404b, respectively. As noted above the subject word file for 404a and object word file 404b are based on extracted Subjects and objects from the core noun phrase file, step 404. These elements have domain spe cific characteristics. WordNet also organizes the word senses in various categories called “Category ID. If for some sub sequence of the predicate, all the verb senses have the same category ID then that category ID is associated with the extracted link. The WordNet category ID is represented in the VerbCategorized.dat file 404c.

For example, relative to exemplary phraseology 402 the word “prosuper” would appear in the subject word file 404a having been extracted from the core noun phrase file. The number of words appearing between subject and predicate;

5

10

15

25

30

35

40

45

50

55

60

65

8 and between predicate and object must be less than or equal to the gap cutoff parameters previously set. Other details are specified in step 430.

Details of the step 134, generating initial ontology from single-sense phrases in a file, are illustrated in the flow dia gram of FIG. 2E of processing 500. Addition of phraseology from the text fragment 202a to the ontology is illustrated in SO2 of FIG. 2E.

In step 504 a determination is made as to whether all entries in the CombinedFile.txt.PhrFreq.csv file, step 218 of FIG.2B have been processed. If so, a determination is made in step 506 as to whether all entries in the instance list have been processed. If so, processing continues in step 136 of FIG. 2A, if not the next entry in the instance list is selected in step 508. In step 510, a determination is made as to whether a term exists in the ontology of which this entry can be an instance. If so, it is added as an instance and processing continues in step 506. Otherwise, processing goes to step 506. When all of the entries in the combined file, step 504 have

not been processed, the next entry is selected for processing, step 516. If the next entry as an instance type in WordNet, step 518, the entry E is added to the instance list step 520. If not, an evaluation is made step 522 as to whether the entry E exists in the ontology. If not, an evaluation is made in step 524 as to the whether the entry E exists in WordNet. If not, in step 526 an evaluation is made as to whether the entry E is the same as its stemmed version. If it is, in step 528 entry E is set to be the stemmed version of E. An evaluation is made in step 530 as to whether entry E has

a single-sense which is a noun. If so the Synset associated with entry E and its parent hierarchy are added to the ontology file as in step 532.

Details of the step 136, add core noun phrase concepts to ontology, are illustrated in the flow diagram of FIG. 2F, pro cessing 600. A determination is made in step 604 as to whether all entries in the core noun phrases file, step from 116, have been processed. If not the next phrase in the file is processed step 606. A determination is made as to whether the phrase already

exists in the ontology, step 608. If not, an evaluation is made as to whether the phrase is a multi-word phrase step 610. If not the phrase is added to list PossConcept, step 612.

If the phrase is a multi-word phrase, each combination of elements P1 P2 are processed where P1 and P2 are separated by a space, step 616. An evaluation is made as to whether element P2 exists in the ontology, step 618. If so, element P1 is associated with P2 in a list named PossOntSubClass, step 620. Otherwise, in step 622 element P1 is associated with P2 in a list named PossConceptSubClass.

In the event that all entries in the file have been processed, step 604, the respective lists PossOntSubClass, PossConcept SubClass and PossConcept are processed respectively in steps 624, 628 and 630. Details of steps 624, 628 and 630 are illustrated respectively in the flow diagrams of FIGS. 2G, 2H, 2I which respectively illustrate processing 700-1, 700-2 and 700-3. The details of the processing of FIGS. 2G, 2H, 2I would be readily understandable by those of skill in the art after a review thereof. No further discussion of those figures is needed.

Subsequent to processing the PossConcept list 630, previ ously extracted links, step 122 are added to the ontology, step 138. Details of the processing of step 138 are illustrated in the flow diagram of FIG. 2J, processing 800. A determination is made, step 804 as to whether any addi

tional entries exist from the link file which need processing. If not, processing continues with step 140.

US 7,987,088 B2

Where there are additional entries, the next entry is selected, step 806. A determination is made as to whether or not the predicate Pred in the entry exists in the ontology, step 808. If not an evaluation is made as to whether the predicate has a single sense which is a verb in WordNet, step 810. If not an instance is created of LinksToReview in the ontology, step 812. If desired a subject matter expert could review the cre ated instance.

Alternately, in step 814 a WordNet word sense is created for each word in the verb synset in the ontology. Finally, in 816 an object property is created for the verb synset and all of its parents in the ontology.

Ontology processing and structuring of the document is completed as the execution of step 138. If desired, the ontol ogy can be reviewed and compared with other ontologies as discussed previously, step 140.

FIGS. 4A and 4B illustrate results of processing a Chronic Cough Management Guideline. The links extracted are shown in FIG. 4A and a portion of the generated ontology is shown in FIG. 4B. These are generated using the processing previously discussed of FIG. 2A. The example of FIG.4A-4B illustrates the flexibility and the domain independent charac teristics of the processing 100 of FIG. 2A.

From the foregoing, it will be observed that numerous variations and modifications may be effected without depart ing from the spirit and scope of the invention. It is to be understood that no limitation with respect to the specific apparatus illustrated herein is intended or should be inferred. It is, of course, intended to cover by the appended claims all Such modifications as fall within the scope of the claims.

The invention claimed is: 1. A domain independent method of creating an ontology

comprising: a programmable processor automatically extracting

phrases from one or more documents independent of a domain of the one or more documents, wherein said extracting of the phrases further comprises separating a portion of a content of at least some of the extracted phrases based upon barrier characters;

the programmable processor extracting core noun phrases from the one or more documents, wherein a plurality of core noun phrases is identified at least in part based on an absence of each of the plurality of core noun phrases from each of an adjective word list, a verb word list, and a barrier word list;

the programmable processor extracting links from the one or more documents based at least in part on the plurality of core noun phrases; and

the programmable processor generating an ontology in accordance with at least the extracted phrases.

2. A method as in claim 1 where generating the ontology includes adding core noun phrases.

3. A method as in claim 1 where generating the ontology includes adding link concepts.

4. A method as in claim 2 where generating the ontology includes adding link concepts.

5. A method as in claim 1 where extracting phrases includes splitting sentences in the documents into fragments.

6. A method as in claim 2 which includes processing frag ments into phrases.

7. A method as in claim 6 which includes limiting phrases to a predetermined number of words.

8. A method as in claim 7 which includes creating a fre quency indicator for each phrase.

9. A method as in claim 8 which includes removing phrases that fail to exceed a predetermined frequency parameter.

10

15

25

30

35

40

45

50

55

60

65

10 10. A method as in claim 1 which includes establishing a

frequency indicator for each of the plurality of core noun phrases.

11. A method as in claim 10 where those core noun phrases whose respective frequency indicator fails to exceed a prede termined threshold are deleted.

12. A method as in claim 1 where extracting the links includes automatically generating Subject and object lists from the one or more documents.

13. A method as in claim 12 which includes forming multi element links that include a Subject element, a predicate ele ment and an object element.

14. A method as in claim 13 where the multi-element links are sorted in accordance with a predetermined criterion.

15. A method as in claim 14 where generating includes evaluating each phrase and, in accordance with a first crite rion, determining if each respective phrase should be added to the ontology.

16. A method as in claim 15 where generating includes evaluation of each core noun phrase, and in accordance with a second criterion determining if each respective core noun phrase should be added to the ontology.

17. A method as in claim 16 where generating includes evaluating each link, and in accordance with a third criterion determining if at least portions of each link should be added to the ontology.

18. An apparatus comprising: a processor including first Software, executed by the pro

cessor, that analyzes a document and forms an extracted phrases file based on barrier characters identified in the document;

the processor including second software, executed by the processor, that analyzes the document and forms a core noun phrases file comprising a plurality of core noun phrases independent of a domain of the document wherein the processor identifies the plurality of core noun phrases at least in part based on an absence of each of the plurality of core noun phrases from each of an adjective word list, a verb word list, and a barrier word list;

the processor including third software, executed by the processor, that analyzes the document and forms a link file based at least in part on the plurality of core noun phrases; and

the processor including fourth software, executed by the processor, that forms an ontology in accordance with Selected phrases in the extracted phrases file.

19. An apparatus as in claim 18 which includes fifth soft ware, executable by the processor, that adds core noun phrase concepts from the core noun phrases file to the ontology.

20. An apparatus as in claim 19 which includes sixth soft ware, executable by the processor, that adds to the ontology link concepts from the link file.

21. An apparatus as in claim 20 which includes a display for reviewing the ontology.

22. An ontology generating system comprising: first Software, recorded on a non-transitory computer read

able medium, that extracts and stores phrases from at least one text Source independent of a domain of the text Source;

second Software, recorded on a non-transitory computer readable medium, that extracts and stores a plurality of core noun phrases from the extracted phrases, wherein the plurality of core noun phrases is identified at least in part based on an absence of each of the plurality of core noun phrases from each of an adjective word list, a verb word list, and a barrier word list;

US 7,987,088 B2 11

third software, recorded on a computer readable medium, that extracts and stores links from the extracted phrases based at least in part on the plurality of core noun phrases; and

fourth Software, recorded on a computer readable medium, that generates an ontology for the at least one text SOUC.

23. A system as in claim 22 where the fourth software generates the ontology Substantially automatically irrespec tive of a subject matter of the at least one text source.

12 24. A system as in claim 22 which includes additional

software for combining textual material from a plurality of Sources to form the at least one text Source.

25. A system as in claim 24 which includes at least one processor which executes at least some of the first, second, third or fourth software.