Towards a Multilingual Medical Lexicon Kornél Markó 1, Robert Baud 2, Pierre Zweigenbaum 3, Magnus Merkel 4, Lars Borin 5, Stefan Schulz 1 1 Freiburg University.

Slides:



Advertisements
Similar presentations
OLIF V2 Gr. Thurmair April OLIF April 2000 OLIF: Overview Rationale Principles Entries Descriptions Header Examples Status.
Advertisements

Using OLIF, The Open Lexicon Interchange Format Susan McCormick OLIF2 Consortium October 1, 2004.
WP 10 Multilingual Access Philipp Daumke, Stefan Schulz.
SemanticMining WP20 meeting Freiburg, March 29 – 20, 2004.
Almen sproglig viden og metode (General Linguistics)
Morphology.
Intelligent Information Retrieval CS 336 –Lecture 3: Text Operations Xiaoyan Li Spring 2006.
Morphology and Lexicon Chapter 3
Example Database English-German Dictionary
Ewa Rudnicka, Wojciech Witkowski, Maciej Piasecki G4.19 Research Group Institute of Informatics, Wrocław University of Technology nlp.pwr.wroc.pl plwordnet.pwr.wroc.pl.
HISA ltd. Biography proforma MEDINFO Lygon Street, Brunswick East 3057 Australia Presenter Name: Stefan Schulz Country:1. Germany, 2. Brazil Qualification(s):
Andrade et al. Corpus-based Error Detection in a Multilingual Medical Thesaurus HISA ltd. Biography proforma MEDINFO Lygon Street, Brunswick East.
Eleni Galiotou, Dept. of Informatics
1 A Hidden Markov Model- Based POS Tagger for Arabic ICS 482 Presentation A Hidden Markov Model- Based POS Tagger for Arabic By Saleh Yousef Al-Hudail.
Morphology Words and Rules. Lexicon collection of the meaningful sound and their meanings in a language dictionaries attempt to be written versions of.
Automatic Lexicon Acquisition for a Medical Cross-Language Information Retrieval System Kornél Markó, Stefan Schulz, Udo Hahn Freiburg University Hospital,
Multilingual Access to Biomedical Documents Stefan Schulz, Philipp Daumke Institute of Medical Biometry and Medical Informatics University Medical Center.
CSE 730 Information Retrieval of Biomedical Data The use of medical lexicon in biomedical IR.
Grammatical frameworks Inflectional morphology. Grammar In the Middle Ages, grammatica […] chiefly meant the knowledge or study of Latin, and were hence.
Getting started with Sanskrit grammar. Inflectional form: Root + Affix = Stem Stem + Inflectional ending = Word.
Kalyani Patel K.S.School of Business Management,Gujarat University.
Evaluating the Contribution of EuroWordNet and Word Sense Disambiguation to Cross-Language Information Retrieval Paul Clough 1 and Mark Stevenson 2 Department.
6th Intex Workshop, Sofia May th Intex Workshop & 10 years of (Silberztein, 1993) Sofia, May 2003.
SemanticMining: Major Events July - December 2005.
Betsy L. Humphreys Betsy L. Humphreys Associate Director for Library Operations NLM, NIH, HHS NLM, NIH, HHS National Library.
The European Thesaurus on International Relations and Area Studies A Multilingual Resource for Indexing, Retrieval, and Translation SWP Michael Kluck and.
Phonemes A phoneme is the smallest phonetic unit in a language that is capable of conveying a distinction in meaning. These units are identified within.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology Automated coding of diagnoses - three methods compared Presenter.
Semi-Automated Extension of a Specialized Medical Lexicon for French Bruno Cartoni & Pierre Zweigenbaum LIMSI-CNRS, France.
Reasons to Study Lexicography  You love words  It can help you evaluate dictionaries  It might make you more sensitive to what dictionaries have in.
© Dialogos Speech Communications S.A. 2 nd Workshop for the Internationalization of SSML Dialogos's Position.
English Review for Final These are the chapters to review. In Textbook: Chapter 1 Nouns Chapter 2 Pronouns Chapter 3 Adjectives Chapter 4 Verbs Chapter.
Using a Lemmatizer to Support the Development and Validation of the Greek WordNet Harry Kornilakis 1, Maria Grigoriadou 1, Eleni Galiotou 1,2, Evangelos.
Copy right 2003 Adam Pease permission to copy granted so long as slides and this notice are not altered Language to Logic Translation.
Integrating Semantic Dictionaries for English, French and Bulgarian into the NooJ System for the Purposes of Information Retrieval Svetla Koeva, Max Silbetztein.
Morphology An Introduction to the Structure of Words Lori Levin and Christian Monson Grammars and Lexicons Fall Term, 2004.
Stefan Schulz, Kornél Markó, Philipp Daumke, Udo Hahn, Susanne Hanser, Percy Nohama, Roosewelt Leite de Andrade, Edson Pacheco, Martin Romacker Semantic.
Linguistics The ninth week. Chapter 3 Morphology  3.1 Introduction  3.2 Morphemes.
SVETLA KOEVA SVETLOZARA LESEVA BORISLAV RIZOV. The project Automatic information extraction based on semantic relations (RILA – a bilateral co-operation.
Sharing Ontologies in the Biomedical Domain Alexa T. McCray National Library of Medicine National Institutes of Health Department of Health & Human Services.
By: Jeremy Pagnotti.  Phonetic language (no silent letters)  No particular word order  Grammatical function of nouns and verbs displayed by endings.
Layered MorphoSaurus Lexicon Extension. Problem Confuse and arbitrary synonym classes of non-medical concepts High ambiguity of general (non- terminological)
Parsing and Translating
Natural Language Processing Chapter 2 : Morphology.
Reference Section. Copyright © Houghton Mifflin Company. All rights reserved.R | 2 1. Personal pronouns.
Leonid Iomdin Institute for Information Transmission Problems, Russian Academy of Sciences
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Automatic Document Indexing in Large Medical Collections.
Annual Review, Brussels March XX, 2006 SemanticMining No Annual Review NoE No Semantic Interoperability and Data Mining in Biomedicine WP20.
A knowledge rich morph analyzer for Marathi derived forms Ashwini Vaidya IIIT Hyderabad.
Inflection. Inflection refers to word formation that does not change category and does not create new lexemes, but rather changes the form of lexemes.
This presentation is intended to discuss about the development and the changes of the English adjectives in Old English, Middle English, and Modern English.
Learning to Generate Complex Morphology for Machine Translation Einat Minkov †, Kristina Toutanova* and Hisami Suzuki* *Microsoft Research † Carnegie Mellon.
Large-Scale Evaluation of a Medical Cross- Language Information Retrieval System Kornél Markó 1,2, Philipp Daumke 1,2, Stefan Schulz 2, Rüdiger Klar 2,
The theory of word classes in modern grammar studies
UNIFIED MEDICAL LANGUAGE SYSTEMS (UMLS)
Multilingual Medical Lexicon
Definite and indefinite articles
Morphology Morphology Morphology Dr. Amal AlSaikhan Morphology.
ENGLISH MORPHOLOGY Week 4.
Basics of the second foreign language theory
Multilingual Biomedical Dictionary
Word Classes and Affixes
Getting started with Sanskrit grammar
Chapter 4 Basics of English Grammar
Words, Wordings & Forewordings
Morphoogle - A Multilingual Interface to a Web Search Engine
Eine and ein: indefinite articles
Agenda diēs Martis, a.d. iii Id. Sept. A.D. MMXVIII
Vocabulary/Lexis LEXIS: n., collective, uncountable
Introduction to English morphology
Presentation transcript:

Towards a Multilingual Medical Lexicon Kornél Markó 1, Robert Baud 2, Pierre Zweigenbaum 3, Magnus Merkel 4, Lars Borin 5, Stefan Schulz 1 1 Freiburg University Hospital, Department of Medical Informatics, Germany 2 University Hospitals of Geneva, Service of Medical Informatics, Switzerland 3 Inserm, U729; Assistance Publique -- Paris Hospitals, STIM; Inalco, CRIM, France 4 Linköping University, Department of Computer and Information Science, Sweden 5 Göteborg University, NLP Section, Sweden

What matters for a multilingual dictionary ? Entries: Base forms + all lexical variants Morpho-syntactic information –POS, number, gender, case, tense, … Coverage –With respect to a given domain Multilingualism –Translation dictionaries

Available Sources Lexical Information CoverageMultilingualism WordNet POS155,000 words General language English EuroWordNet POS30,000-50,000 words General language Dutch, Italian, Spanish, German, French, Czech, Estonian, … EuroDicAutum 630,000 words and phrases EU activities, finance, agriculture, legislation, transport,… Dutch, French, German, Italian, Danish, English, Greek, Portuguese, Spanish, Finnish, Swedish UMLS Metathesaurus - > 1M words and phrases Biomedicine English, German, French, Spanish, French, Swedish, Russian, … UMLS Specialist Lexicon POS, number, case, tense, person, inflection types, … 250,000 words and acronyms Biomedicine English

Multilingual Lexicon Creation 1.Create/Collect available (monolingual) resources 2.Define an interchange format 3.Convert resources to interchange format 4.Define a cross-lingual linking format 5.Define criteria for semantic linkage 6.Semantically align lexical entries

Multilingual Lexicon Creation 1.Create/Collect available (monolingual) resources 2.Define an interchange format 3.Convert resources to interchange format 4.Define a cross-lingual linking format 5.Define criteria for semantic linkage 6.Semantically align lexical entries

Collect Monolingual Lexical Resources with full lexical information French UMLF lexicon from different French health-related organizations and the University Hospitals of Geneva, Switzerland (33,718 entries) English medical lexicon from Linköping University, Sweden (22,686 entries) Swedish medical lexicon from Linköping University (23,223 entries) Swedish medical lexicon from Göteborg University, Sweden (6,786 entries) German Specialist Lexicon from Freiburg University Hospital, Germany (41,316 entries) English Specialist Lexicon, which is part of the UMLS (96,621 entries, avoiding acronyms and chemical names) A total of 224,351 entries

Collect Monolingual Lexical Resources with full lexical information French UMLF lexicon from different French health-related organizations and the University Hospitals of Geneva, Switzerland (33,718 entries) English medical lexicon from Linköping University, Sweden (22,686 entries) Swedish medical lexicon from Linköping University (23,223 entries) Swedish medical lexicon from Göteborg University, Sweden (6,786 entries) German Specialist Lexicon from Freiburg University Hospital, Germany (41,316 entries) English Specialist Lexicon, which is part of the UMLS (96,621 entries, avoiding acronyms and chemical names) A total of 224,351 entries Large variability in scope and granularity

Multilingual Lexicon Creation Collect available (monolingual) resources Define an interchange format Convert resources to that format Define a cross-lingual linking format Define criteria for semantic linkage Semantically align lexical entries

Lexicon Interchange Format (cf. Baud et al. AMIA 05) FieldDescriptionDefinition IDUnique Identifier FrmInflected FormInflection of Lem MfrMorphosyntactic features of inflected formCoded in MULTEXT LemLemmaThe basic form of the entry MulMorpho-syntactic featuresCoded in MULTEXT LngLanguageEN,DE,FR,SW TypEntry TypeBasic entry (B) Subword entry (S) Compound entry (C) Term entry (T) PrtDecompositionParts of a compound entry StrHeadHead of a compound/term entry RefReference lemmaID of lemma‘s entry

MULTEXT Standard Part-of-Speech =============== Noun N Verb V Adjective A Adverb R … Nouns (N) = ======= ============= ==== P ATT VAL C = ======= ============= ==== 1 Type common c proper p Gender masculine m feminine f neuter n Number singular s plural p Case nominative n genitive g dative d accusative a Verbs (V) = ============ ============= ==== P ATT VAL C = ============ ============= ==== 1 Type main m auxiliary a modal o Mood/VForm indicative i subjunctive s imperative m conditional c infinitive n participle p gerund g supine s base b Tense present p imperfect i future f past s Person first 1 second 2 third Number singular s plural p Gender masculine m feminine f neuter n Adjectives (A) = ============ =============== ==== P ATT VAL C = ============ =============== ==== 1 indefinite i possessive s Degree positive p comparative c superlative s Gender masculine m feminine f neuter n Number singular s plural p Case nominative n genitive g dative d accusative a Articles (T) = ============ =============== ==== P ATT VAL C = ============ =============== ==== 1 Type definite d indefinite i Gender masculine m feminine f neuter n Number singular s plural p

MULTEXT Standard Part-of-Speech =============== Noun N Verb V Adjective A Adverb R … Nouns (N) = ======= ============= ==== P ATT VAL C = ======= ============= ==== 1 Type common c proper p Gender masculine m feminine f neuter n Number singular s plural p Case nominative n genitive g dative d accusative a Verbs (V) = ============ ============= ==== P ATT VAL C = ============ ============= ==== 1 Type main m auxiliary a modal o Mood/VForm indicative i subjunctive s imperative m conditional c infinitive n participle p gerund g supine s base b Tense present p imperfect i future f past s Person first 1 second 2 third Number singular s plural p Gender masculine m feminine f neuter n Adjectives (A) = ============ =============== ==== P ATT VAL C = ============ =============== ==== 1 indefinite i possessive s Degree positive p comparative c superlative s Gender masculine m feminine f neuter n Number singular s plural p Case nominative n genitive g dative d accusative a Articles (T) = ============ =============== ==== P ATT VAL C = ============ =============== ==== 1 Type definite d indefinite i Gender masculine m feminine f neuter n Number singular s plural p LemmaMULTEXT nail (EN)Nc-sn doigt (FR)Ncmsn digital (SW)A-pfsn

Multilingual Lexicon Creation 1.Collect available (monolingual) resources 2.Define an interchange format 3.Convert resources to interchange format 4.Define a cross-lingual linking format 5.Define criteria for semantic linkage 6.Semantically align lexical entries

Monolingual Resources EN|DIM:20501|B||abdomen|Ncns||||||||| EN|DIM:20502|B||abdominal|Afpns||||||||| EN|DIM:20503|B||abduction|Ncns||||||||| EN|DIM:20504|B||abductor|Ncns||||||||| SV|LIU_SV142_A|B||buk|nc0sn|||||||||| SV|LIU_SV143_A|B||abdominal|afpusn|||||||||| SV|LIU_SV144_A|T||från buk|nc0sn||||||buk|||| SV|LIU_SV145_A|C||bukabscess|nc0sn|||||buk__abscess||||| SV|LIU_SV146_A|T||abdominell aktinomykos|nc0sn||||||aktinomykos|||| SV|LIU_SV147_A|C||bukaorta|nc0sn|||||buk__aorta||||| DE|UKLFR:1045|B||Atrophie|Ncfsn|||19|||||| DE|UKLFR:1046|B||Atropin|Ncnsn|||11|||||| DE|UKLFR:1047|B||Atropinvergiftung|Ncfsn|||20|||||| DE|UKLFR:1048|B||Attacke|Ncfsn|||19|||||| DE|UKLFR:1049|B||Attest|Ncnsn|||11|||||| FR|UMLF:10012|B||distiller|Vmn||||||||| FR|UMLF:10013|B||distinct|Afms||||||||| FR|UMLF:10014|B||distinguer|Vmn||||||||| FR|UMLF:10015|C||distocclusion|Ncfs|||||dist__occlusion|occlusion|||

POS: Common Noun EN|DIM:20501|B||abdomen|Ncns||||||||| EN|DIM:20502|B||abdominal|Afpns||||||||| EN|DIM:20503|B||abduction|Ncns||||||||| EN|DIM:20504|B||abductor|Ncns||||||||| SV|LIU_SV142_A|B||buk|nc0sn|||||||||| SV|LIU_SV143_A|B||abdominal|afpusn|||||||||| SV|LIU_SV144_A|T||från buk|nc0sn||||||buk|||| SV|LIU_SV145_A|C||bukabscess|nc0sn|||||buk__abscess||||| SV|LIU_SV146_A|T||abdominell aktinomykos|nc0sn||||||aktinomykos|||| SV|LIU_SV147_A|C||bukaorta|nc0sn|||||buk__aorta||||| DE|UKLFR:1045|B||Atrophie|Ncfsn|||19|||||| DE|UKLFR:1046|B||Atropin|Ncnsn|||11|||||| DE|UKLFR:1047|B||Atropinvergiftung|Ncfsn|||20|||||| DE|UKLFR:1048|B||Attacke|Ncfsn|||19|||||| DE|UKLFR:1049|B||Attest|Ncnsn|||11|||||| FR|UMLF:10012|B||distiller|Vmn||||||||| FR|UMLF:10013|B||distinct|Afms||||||||| FR|UMLF:10014|B||distinguer|Vmn||||||||| FR|UMLF:10015|C||distocclusion|Ncfs|||||dist__occlusion|occlusion|||

POS: Adjective EN|DIM:20501|B||abdomen|Ncns||||||||| EN|DIM:20502|B||abdominal|Afpns||||||||| EN|DIM:20503|B||abduction|Ncns||||||||| EN|DIM:20504|B||abductor|Ncns||||||||| SV|LIU_SV142_A|B||buk|nc0sn|||||||||| SV|LIU_SV143_A|B||abdominal|afpusn|||||||||| SV|LIU_SV144_A|T||från buk|nc0sn||||||buk|||| SV|LIU_SV145_A|C||bukabscess|nc0sn|||||buk__abscess||||| SV|LIU_SV146_A|T||abdominell aktinomykos|nc0sn||||||aktinomykos|||| SV|LIU_SV147_A|C||bukaorta|nc0sn|||||buk__aorta||||| DE|UKLFR:1045|B||Atrophie|Ncfsn|||19|||||| DE|UKLFR:1046|B||Atropin|Ncnsn|||11|||||| DE|UKLFR:1047|B||Atropinvergiftung|Ncfsn|||20|||||| DE|UKLFR:1048|B||Attacke|Ncfsn|||19|||||| DE|UKLFR:1049|B||Attest|Ncnsn|||11|||||| FR|UMLF:10012|B||distiller|Vmn||||||||| FR|UMLF:10013|B||distinct|Afms||||||||| FR|UMLF:10014|B||distinguer|Vmn||||||||| FR|UMLF:10015|C||distocclusion|Ncfs|||||dist__occlusion|occlusion|||

POS: Compound EN|DIM:20501|B||abdomen|Ncns||||||||| EN|DIM:20502|B||abdominal|Afpns||||||||| EN|DIM:20503|B||abduction|Ncns||||||||| EN|DIM:20504|B||abductor|Ncns||||||||| SV|LIU_SV142_A|B||buk|nc0sn|||||||||| SV|LIU_SV143_A|B||abdominal|afpusn|||||||||| SV|LIU_SV144_A|T||från buk|nc0sn||||||buk|||| SV|LIU_SV145_A|C||bukabscess|nc0sn|||||buk__abscess||||| SV|LIU_SV146_A|T||abdominell aktinomykos|nc0sn||||||aktinomykos|||| SV|LIU_SV147_A|C||bukaorta|nc0sn|||||buk__aorta||||| DE|UKLFR:1045|B||Atrophie|Ncfsn|||19|||||| DE|UKLFR:1046|B||Atropin|Ncnsn|||11|||||| DE|UKLFR:1047|B||Atropinvergiftung|Ncfsn|||20|||||| DE|UKLFR:1048|B||Attacke|Ncfsn|||19|||||| DE|UKLFR:1049|B||Attest|Ncnsn|||11|||||| FR|UMLF:10012|B||distiller|Vmn||||||||| FR|UMLF:10013|B||distinct|Afms||||||||| FR|UMLF:10014|B||distinguer|Vmn||||||||| FR|UMLF:10015| C||distocclusion|Ncfs|||||dist__occlusion|occlusion|||

Multi Word Nouns EN|DIM:20501|B||abdomen|Ncns||||||||| EN|DIM:20502|B||abdominal|Afpns||||||||| EN|DIM:20503|B||abduction|Ncns||||||||| EN|DIM:20504|B||abductor|Ncns||||||||| SV|LIU_SV142_A|B||buk|nc0sn|||||||||| SV|LIU_SV143_A|B||abdominal|afpusn|||||||||| SV|LIU_SV144_A|T||från buk|nc0sn||||||buk|||| SV|LIU_SV145_A|C||bukabscess|nc0sn|||||buk__abscess||||| SV|LIU_SV146_A|T||abdominell aktinomykos|nc0sn||||||aktinomykos|||| SV|LIU_SV147_A|C||bukaorta|nc0sn|||||buk__aorta||||| DE|UKLFR:1045|B||Atrophie|Ncfsn|||19|||||| DE|UKLFR:1046|B||Atropin|Ncnsn|||11|||||| DE|UKLFR:1047|B||Atropinvergiftung|Ncfsn|||20|||||| DE|UKLFR:1048|B||Attacke|Ncfsn|||19|||||| DE|UKLFR:1049|B||Attest|Ncnsn|||11|||||| FR|UMLF:10012|B||distiller|Vmn||||||||| FR|UMLF:10013|B||distinct|Afms||||||||| FR|UMLF:10014|B||distinguer|Vmn||||||||| FR|UMLF:10015|C||distocclusion|Ncfs|||||dist__occlusion|occlusion|||

Multilingual Lexicon Creation 1.Collect available (monolingual) resources 2.Define an interchange format 3.Convert resources to interchange format 4.Define a cross-lingual linking format 5.Define criteria for semantic linkage 6.Semantically align lexical entries

Linking Format FieldDescriptionDefinition SrcSourceSource ID to be linked with target entry TarTargetTarget ID to be linked with source entry LnkType of Relation??? Relation Type: Language-specific characteristics of number, case, gender Multiple derivations, e.g. attributive or predicative adjectives, definite or indefinite objects Links represent possible translations.

Multilingual Lexicon Creation 1.Collect available (monolingual) resources 2.Define an interchange format 3.Convert resources to interchange format 4.Define a cross-lingual linking format 5.Define criteria for semantic linkage 6.Semantically align lexical entries

Simple Relation Types REL1 –A and B share the same POS and MULTEXT features REL2 –A and B share the same POS, but at least one MULTEXT feature differs REL3 –A and B do not share same POS

Multilingual Lexicon Creation 1.Collect available (monolingual) resources 2.Define an interchange format 3.Convert resources to interchange format 4.Define a cross-lingual linking format 5.Define criteria for semantic linkage 6.Semantically align lexical entries

Linking using Subword Indexing MorphoSaurus engine extracts meaningful subwords from medical text Maps them to language-independent concept IDs Covers English, German, Portuguese, French, Spanish, Swedish Validated for cross-lingual text retrieval

(MIDs)

Semantic Enrichment All inflected forms (Frm-field, if available) or base forms (Lem-field) are processed with MorphoSaurus. Resulting representations are added to lexicon entries: –… –EN|DIM:20501|B||abdomen|Ncns|||||||||| #abdom –DE|UKLFR:1383|B||Bauch|Ncmsn|||1u||||||| #abdom –…

Linking Algorithm FOR EVERY lexeme i and its attributes in the list DO FOR EVERY lexeme j (starting from i) and its attributes in the list DO IF MorphoSaurus-Representation of input is identical THEN IF POS and MUL-values of i and j are identical THEN type = REL1 # synonymy/translation IF POS-values are identical, but not MUL-values THEN type = REL2 IF POS-values differ THEN type = REL 3 # derivation PRINT “ID(i) | ID(j) | type”

Results Language Pair# RelationsDifferent Lemmas English-German126,50431,544 English-French70,68024,368 English-Swedish86,65534,030 French-Swedish21,6048,312 French-German32,65910,458 German-Swedish41,46912,105 TOTAL379,571120,817 Additional 271,971 intralingual synonymy relationships found (sums up to 651,542) REL1 (10%), REL2 (44%), REL3 (46%)

Preliminary Evaluation: Coverage Most comprehensive resource for medical terminology: UMLS Metathesaurus How many terms are covered? How many relations are covered?

Coverage: Terms LanguageUMLS *CoveredSynonyms**Additional** English122,03532,668 (27%)3,80768,842 German21,1622,832 (13%)1,26923,379 French10,2603,590 (35%)30925,923 Swedish12,0128,520 (71%)99417,579 TOTAL165,469189,712 * Only single-word preferred entries ** only REL1 and REL2 (no derivation) UMLSMultilingual Lexicon

Coverage: Relations LanguageUMLSCoveredSynonyms*Additional* English-German15,9791,259 (8%)8,80121,484 English-French12,5891,783 (14%)6,97415,611 English-Swedish9,5543,403 (36%)10,12420,503 German-French9, (9%)7738,835 German-Swedish10, (8%)1,6999,596 French-Swedish6,7931,109 (16%)1,9115,292 TOTAL64,837120,817 * only REL1 and REL2

Conclusion Framework for integrating heterogenous lexical resources Integration of English, German, French and Swedish sources Simple Linkage Format for coding lexical relations Generating substantial amount of synonym mappings and translations using MorphoSaurus subword indexing Partly validated using UMLS Metathesaurus Extensive evaluation is ongoing (presentation at MEDINFO)

Morphosaurus Resources Subword-Lexicon: –Organizes subwords in several languages (English, German, Portuguese, Spanish, Fench, Swedish) Subword-Thesaurus: –Groups synonymous subwords (within and between languages) Subword-Segmeter: –Extraction of Subwords and Assignment of Equivalence Classes Morphosaurus- Identifier (MID) Morphosaurus

Subword Approach Subwords are atomic, conceptual or linguistic units: –Stems: stomach, gastr, diaphys –Prefixes: anti-, bi-, hyper- –Suffixes: -ary, -ion, -itis –Infixes: -o-, -s- Equivalence classes contain synonymous subwords and their translations: #derma = { derm, cutis, skin, haut, kutis, pele, cutis, piel, … } #inflamm = { inflam, -itic, -itis, entzuend, -itis, -itisch, inflam, flog, inflam, flog, -iolitis,... }

Subwords: Lexicon & Thesaurus Subword Lexicon: gastr stomach magen ventric chamber hepat,hepar liver leber nephr ren kidney nier Subword Thesaurus: groups synonymous subwords to equivalence classes #GASTR #CHAMBER #HEPAR #NEPHR

Example high tsh value s suggest the diagnos is of primar y hypo thyroid ism er hoeh te tsh wert e erlaub en die diagnos e einer primaer en schilddruese n unter funktion Segmenter Subword-Lexicon High TSH values suggest the diagnosis of primary hypo- thyroidism... Original Erhöhte TSH-Werte erlauben die Diagnose einer primären Schilddrüsenunterfunktion... high tsh values suggest the diagnosis of primary hypo- thyroidism... erhoehte tsh werte erlauben die diagnose einer primaeren schilddruesenunterfunktion... Orthografic Rules Orthographic Normalization #up tsh #value #suggest #diagnost #primar #small #thyre Interlingua #up tsh #value #permit #diagnost #primar #thyre #hypo #function Subword- Thesaurus Semantic Normalization

Example high tsh value s suggest the diagnos is of primar y hypo thyroid ism er hoeh te tsh wert e erlaub en die diagnos e einer primaer en schilddruese n unter funktion Segmenter Subword-Lexicon High TSH values suggest the diagnosis of primary hypo- thyroidism... Original Erhöhte TSH-Werte erlauben die Diagnose einer primären Schilddrüsenunterfunktion... high tsh values suggest the diagnosis of primary hypo- thyroidism... erhoehte tsh werte erlauben die diagnose einer primaeren schilddruesenunterfunktion... Orthografic Rules Orthographic Normalization #up tsh #value #suggest #diagnost #primar #small #thyre Interlingua #up tsh #value #permit #diagnost #primar #thyre #hypo #function Subword- Thesaurus Semantic Normalization