Integration of structured museum descriptions using domain ontology and thesauri
Speaker: Vladimir Ivanov, Kazan State University, Russia
Date: 14 May 2008 Time: 14:00-15:30
Location: STEP-C Seminar Room, Building B (1st floor), FORTH, Heraklion, Crete
Host: M.Doerr


Problems of heterogeneous information integration are well known in IT domain. Such problems appear in cultural heritage domain when one tries to represent collections belong to different organizations, which use different data structures. In this presentation I will discuss technical background of those problems in cultural heritage domain. Common mapping process is based on using ontologies in order to represent knowledge encoded in different structures and on using modern natural language processing techniques to deal with terminological diversity. Implemented approach is explained by giving examples from museum databases.

Key points of a detailed presentation (about 1 hour)

  1. Information integration, mapping problems
    • Overview of well-known frameworks, methods, platforms, etc.
    • Ontologies and thesauri in information integration
  2. Integrating museum descriptions
    • Structured vs. unstructured descriptions
    • Heterogeneity in museum databases (examples from Russian museums)
    • Results of the integration process: virtual vs. materialized
    • Using CRM and thesauri in integration
      • Basics: extending CRM with thesaurus terms, interpretation of database schema, intended result - data warehouse with the CRM structure and links to the thesaurus
      • Defining mappings. Mapping ontology
      • Mapping execution. Thesaurus-based conceptual indexing
      • Improving mapping definition. LSA-based method for interpretation the meaning of a DB schema elements
      • Mapping validation
    • Implementation of a common mapping process
      • Lifting database schema (reduction to Description Logics formalism)
      • Classification of structural elements of the schema
      • Mapping definition
      • Execute mapping and put results into a CRM-compliant structure
      • Mapping result validation
      • Examples: mapping databases of 3 Russian museums to the CRM
    • Applications: Natural language query processing
      • Basic algorithm
      • Examples: comparing results to Google search


© Copyright 2007 FOUNDATION FOR RESEARCH & TECHNOLOGY - HELLAS, All rights reserved.