University of Economics, Prague Czech Republic - PowerPoint PPT Presentation

1 / 12
About This Presentation
Title:

University of Economics, Prague Czech Republic

Description:

University of Economics, Prague (Czech Republic) OntoWeb partner no. 63 ... Induction of Probabilistic Models from Patient Records (aiming at MG critiquing) RAINBOW ... – PowerPoint PPT presentation

Number of Views:293
Avg rating:3.0/5.0
Slides: 13
Provided by: nob896
Category:

less

Transcript and Presenter's Notes

Title: University of Economics, Prague Czech Republic


1
University of Economics, Prague (Czech Republic)
  • OntoWeb partner no. 63
  • Contact Vojtech Svátek svatek_at_vse.cz

2
University of Economics, Prague
  • 6th largest
  • in the country
  • 14 thousand students
  • 6 faculties

3
OntoWeb participation
  • Joint workgroup of
  • Department of Information and Knowledge
    Engineering (DIKE)
  • Laboratory for Intelligent Systems, Prague (LISP)
  • at the Faculty of Informatics and Statistics
  • V.Svátek, J.Rauch, P.Berka, V.Sklenák ( PhD
    students and other people)

4
Research Interests
  • Knowledge Modelling and Ontologies
  • Information Science, Information Retrieval,
    Semantic Document Mark-Up
  • Knowledge Discovery in Databases
  • Uncertainty Processing, Bayesian Networks

5
Selected Research Projects
  • IC15 CT 98-0315 Medical Guideline Technology
    (1998-2001)
  • RAINBOW Reusable Architecture for INformation
    Brokering On the Web (2000-?)
  • ARROW Analytical Reports Reusable Over the Web
    (2001-?)

6
Medical Guideline Technology
http//frost.open.ac.uk/mgt/ Formalisation
and Computational Processing of Medical
Guidelines (MG)
  • IST Fourth Framework Programme
  • Project Co-ordinator Z.Zdrahal, Open University
    (UK)
  • Scientific Co-ordinator J.Zvarova, Czech Academy
    of Science
  • UEP R.Jirouek, V.Svátek, et al.

7
Medical Guideline Technology
  • UEP contributes in the areas of
  • Designing a Medical Guideline Ontology
  • Step-by-step mark-up of MG Documents
    (methodology, languages and tools)
  • Modelling the compliance with MG, using the OCML
    Language Data Mining aspects
  • Induction of Probabilistic Models from Patient
    Records (aiming at MG critiquing)

8
RAINBOW http//nb.vse.cz/svatek/rainbow
  • Informal project undertaken at the UEP
    (Co-ordination V.Svátek, P.Berka, V.Sklenák)
  • Aiming at a reusable architecture configurable
    for various tasks of WWW analysis and access

9
RAINBOW
  • Automated extraction of meta-data based on shared
    ontology
  • Labour division wrt. data type
  • natural language, HTML tagging and generic page
    structure, frequencies of terms and concepts, URL
    addressing, link topology, explicit meta-data,
    image information
  • Distributed skeletal planning

10
Analytical Reports Reusable Over the Web
  • Co-ordinator J.Rauch
  • Backround KDD research at the UEP
  • partly within the Solomon European Network
    Project (IST-1999-11495)
  • original KDD system LISP Miner mining
    (statistical) association rules, applications in
    finance and medicine
  • organisation of PKDD99 (J.Rauch as PC Co-Chair)
    and Discovery Challenge 1999-2001 (P.Berka)

11
Analytical Reports Reusable Over the Web
  • Need for reuse of both (template) KDD task
    settings and KDD results
  • Envisaged tools
  • KDD Ontology derived from CRISP-DM, combined with
    domain (e.g. medical) ontologies
  • Analytical Reports as (semi-)automatically
    generated text documents enriched with logical
    formulae extracted from data
  • knowledge (formulae) indexing for automated
    retrieval and reuse over the WWW

12
Thank you for your attention
  • Further questions
  • svatek_at_vse.cz
Write a Comment
User Comments (0)
About PowerShow.com