medis praha 18 11 2005 - masarykova univerzitageotez.geogr.muni.cz/pdf/hadrbolec_medis.pdf ·...
TRANSCRIPT
05.12.2005| Folie 3
Overview
GoalsBirds EyeComponents
TEDGSBLBabelFishGoogleIndexing
Proof of ConceptLessons learned / Conclusions
05.12.2005| Folie 4
Goals 1/2
Creation of an overview of types of disasters with concentration on the target group industry Structuring and implementing of a prototype portal with entrance to information to the disaster managementRealization of the portal surface with a password-protected rangeInstallation of an indexation mechanism for rapid finding of contents from selected Web sites. Analysis and selection from data sources referring to dangerous materials
05.12.2005| Folie 5
Goals 2/2
Programming of an interface to the GSBL (common material data pool of the federal countries, Germany and the German Environment Ministry)Analysis and selection of multilingual glossaries and thesauri in the field of emergency and disaster managementStructure of the thesaurus on Emergencies and Disasters (TED) with more than 1.800 terms in the languages German, English and ItalianLink of the search functions of the Web portal with the thesaurus, in order to be able to look at the same time for synonyms and/or super ordinate terms.
05.12.2005| Folie 7
Birds Eye (Medis Principle) 2/2
Approach:Select multiple existing source and implement it as proof of Concept.
05.12.2005| Folie 8
Proof of Concept
Medis is based on the following components
TED (Thesaurus on Emergencies and Disasters)GSBL (Gemeinsamer StoffdatenpoolBund/Länder [Germany])BabelFish (Translating Service)Google (Search engine)Indexing Engine (Apache Lucene)
05.12.2005| Folie 9
TED 1/2
Sources•Australian Emergency Management Terms Thesaurus (Emergency Management Australia, EMA)
•GEMET General Multi-Lingual Environmental Thesaurus (EEA)
•Eurodicautom (European Terminology Database, EC)
•Brochures of Ministry of interior
•Glossaries: Glossar von Termini derRisikoanalyse, BGVV, Berlin, Safety Glossary, E-D, IAEA, Wien, …
•…
05.12.2005| Folie 11
Gemeinsamer Stoffdatenpool Bund/Länder (GSBL) 1/4
The current GSBL data inventory combines data from the “hazardous substances rapid information” (GSA), the chemicals information system (CHEMIS), the information and communications system for hazardous/ environmentally relevant substances in North Rhine Westphalia (IGS), the hazardous substances database of the states (GDL), from the oncall service and initial deployment information system RESY, data from the Beilstein Institute and fire service specific suggestions for action of the Institute of the Fire Service in Heyrothsberge as well as transport provider related data of the Federal Institute for Materials Research and Testing (BAM).
05.12.2005| Folie 15
BabelFish 1/2
The Babel Fish is small, yellow, and simultaneously translates from one
spoken language to another. Hitchhikers guide to the galaxy
(Douglas Adams)
Problem
• Not everybody knows every word in a different language especially under pressure.
• A resource is only in a different language available. Translate it into a language of the users choice.
05.12.2005| Folie 16
BabelFish 2/2
SolutionBased on a SOAP based API to AltaVista BabelFish translations between different languages (among 19 pairs of languages) are possible.
05.12.2005| Folie 17
This API was used in MEDIS.
With the Google Web APIs service, software developers can query billions of web pages directly from their own computer programs. http://www.google.com/apis/
05.12.2005| Folie 18
Indexing Engine based on ApacheLucene1/2
Problem:
Index dedicated quality assured documents and or parts of an website
which contains relevant information.
05.12.2005| Folie 19
Indexing Engine based on Apache Lucene2/2
Solution1. Develop a WebCrawler and2. index each identified document 3.-n Search (Query index) Display result
05.12.2005| Folie 22
Authentication
Only authorized users should access the content.
1. Authenticate2. Authorize
05.12.2005| Folie 25
Example
1. Search in Thesaurus in your own language
2. Translate into a language of your choice
3. Search parallel in multiple quality assured data sources
Get results in a unified way