=Paper=
{{Paper
|id=Vol-1937/paper5
|storemode=property
|title=Project Coli-conc: Mapping Library Knowledge Organisation Systems
|pdfUrl=https://ceur-ws.org/Vol-1937/paper5.pdf
|volume=Vol-1937
|authors=Uma Balakrishnan,Morsheda Akter
|dblpUrl=https://dblp.org/rec/conf/ercimdl/BalakrishnanA17
}}
==Project Coli-conc: Mapping Library Knowledge Organisation Systems==
https://ceur-ws.org/Vol-1937/paper5.pdf
Project Coli-conc: Mapping Library Knowledge
Organisation Systems
Uma Balakrishnan, Morsheda Akter.
Verbundzentrale der GBV, Germany
Abstract: Knowledge organisation systems (KOS) across libraries in German speaking
countries have always struggled with a lack of homogeneity and consistent subject indexing.
Various tools have been developed to map different KOS, either manually or through
algorithms which allow for automatic or semi-automatic mapping 1 , 2 , 3 . Considering the
complexity and variety of KOS, there is a need for an overarching system which facilitates and
promotes consistent subject indexing and enhances access to information and use of KOS. The
first step towards designing and deploying such a system was the conversion of various KOS
data from their proprietary formats to a uniform format. To meet this requirement, Coli-conc
developed in the pilot phase of the project the JSKOS format 4. The newly created format is
based on SKOS and JSON LD; it eases the use of KOS in web applications and allows more
flexibility in content display of the KOS data. To accelerate and increase the efficiency of the
intellectual mapping process, the project proposes the mapping tool Cocoda that adopts a
dashboard design. This approach gathers data of the KOS that are being mapped with each
other from different sources, consolidates the same and displays it in a concise manner on a
single screen for at-a glance-monitoring. The web-based mapping tool, Cocoda, besides
enabling term/caption or notation search from a source or a target KOS; presents the
hierarchical structure of a queried term/class; and permits browsing of the higher order concepts
of the selected KOS. Additionally, the tool automatically generates mappings of a selected
term/notation, gives options to edit and save the newly created mappings in the centralized
VZG Mapping database integrating a feature to export them in JSKOS and various other
formats. Furthermore, the Coli-conc infrastructure includes a KOS-Registry, a platform for
Concordances with a web interface, JSKOS API for KOS and mappings which are provided as
stand-alone services.
1
Walter, A.-K., Mayr, P., Petras, V., Baerisch, S. (2007)
2
Pfeffer, M. (2013)
3
Lauser, B., Johannsen, G., Caracciolo, C., Keizer, J., van Hage, W. R., Mayr, P. (2008)
4
Voß, J. (2016)
Features of the Coli-conc and technical specifications
1. Coli-conc System Architecture
The system offers an infrastructure to support semi-automatic creation of mappings.
Furthermore, it facilitates the use and exchange of KOS and their mappings. The
application is modular designed and consists of three core parts:
Coli-conc Web (CCWeb)
Data Converter
JSKOS API for KOS and Mappings (KOS-API, KK-API)
Database Server
CCWeb Cocoda
Sources KOS-
Representation
KK-Suggest
KOS-API
KK-Measure
Web
Interface
KK-API
Concordance
s
Kos-Registry
Converter &
Uploader
User
Administration
Database Server
Figure 1: Coli-Conc System Architecture
1.1. CCWeb
The CCWeb is the heart of the system architecture. It comprises of:
The Mapping tool – Cocoda
KOS-Registry
Concordances
User Admin module and a Web Interface
Web
Interface
Presentation
CCWeb Application
Concordance
s
Database Server
KOS-Registry
Database
Cocoda
User
Administration
Figure 2: CCWeb
1.1.1. Mapping Tool – Cocoda
Cocoda is designed to perform multiple tasks to speed up the intellectual mapping
building process between library KOS. The main components of the tool are:
KOS-Representation Module: to search, browse and display hierarchical
intra-KOS structural and content information for concept clarification.
KK-Suggest Module: It combines several techniques and workflows into the
framework (statistical co-occurrences, direct concept mapping and retrieval
of stored mappings from the concordance database) to maximize the results
and generate mapping candidates for a selected source and target KOS.
KK-Measure Module: to monitor and take care of quality assessment of the
KOS and their mappings.
Figure 3: Mapping Tool – Cocoda
1.1.2. KOS-Registry
The KOS-Registry holds a collection of library KOS that are actively in use in the
German speaking countries.5 The records have been classified under various types and
also enriched with metadata based on the NKOS AP 6 format. An interface enables a
keyword search or selection of a specific type of KOS through a drop-down menu and
retrieval of the metadata of each KOS in the selected set. The application is equipped
with an export function in different formats (XLS, JSON and JSKOS). 7 A script
continually updates the registry through automatic data acquisition from BARTOC8.
Figure 4: KOS-Registry
5
Balakrishnan, U., Agne, J.M. (2016)
6
NKOS AP: http://nkos.slis.kent.edu/nkos-ap.html
7
Voß, J., Ledl, A., Balakrishnan, U. (2016)
8
BARTOC: https://bartoc.org/
1.1.3. Concordances
To store, manage, and access mappings of the Coli-conc and other related projects as
well as to integrate the same into the KK-Suggest module, the project has developed a
concordance platform as part of the CCWeb. The database of the platform contains
currently over 200.000 mappings and will be further built up with the help of partner
institutions.
Figure 5: Concordances
2. Technologies and Frameworks deployed for CCWeb
Technologies:
Web Framework:
Servers: JavaServer
Java Jersey Apache Faces (JSF) JavaScript
bundle Tomcat v
7.0 or v 2.2
higher
Application
Servers: Database User Server
J2EE (Java 2
IDE: Servers: Interface: Request
Platform, eclipse Mongodb PrimeFaces Forms:
Enterprise v 5.0
Edition) v 3.2.1 XHTML
Figure 6: Technologies and frameworks for CCWeb
3. Data Conversion
One of the main components of the Coli-conc system is the JSKOS format that was
modelled specifically for the project. Application for DDC (Dewey Decimal
Classification) and RVK (Regensburg Classification) conversion from their
proprietary formats into JSKOS has been developed using MARC4J package9. For
other KOS (Basic Classification and Allgemeine Systematik für Öffentliche
Bibliotheken,…) and mappings the mc2skos Python script10 was extended to support
JSKOS. The conversion scripts for mappings are available at:
https://github.com/gbv/cocoda-mappings.
Figure 7: DDC Marc21 XML data converter Sequence Diagram
9
MARC4J package: http://svn.k-int.com/default/components/marc4j/tutorial.html
10
mc2skos Python script : https://pypi.python.org/pypi/mc2skos
4. JSKOS API
An additional special feature of the Coli-conc architecture is the JSKOS API. Among
the key objectives of the project is to provide uniform and easy access to KOS and
their mappings on the web. This has been affected by the creation of the JSKOS API.
The service has been so far implemented as a database application for DDC, RVK,
Basic classification (BC) and as wrappers to access Gemeinsame Normdatei (GND)
Wikidata, Open Research and Contribution ID (ORCID). However, the use of the
DDC API is subject to a license requirement.
Figure 8: DDC- JSKOS API
Project Partner Institutions
The project is being funded by the German Research Foundation. It has received
support from the German National Library, several expert groups, large academic
libraries and international institutions.
Figure 9: Project Partner Institutions
References
Balakrishnan, U., Agne, J.M. (2016). Coli-conc Survey - Ergebnisse der Online-
Umfrage zur Sacherschließung und Konkordanzprojekten. Retrieved from http://coli-
conc.gbv.de/publications/
Lauser, B., Johannsen, G., Caracciolo, C., Keizer, J., van Hage, W. R., Mayr, P.
(2008). Comparing human and automatic thesaurus mapping approaches in the
agricultural domain. 10th International Conference on Dublin Core and Metadata
Applications. Retrieved from https://arxiv.org/abs/0808.2246
Pfeffer, M. (2013): Automatic creation of mappings between classification systems.
Workshop Klassifikation und Sacherschließung (LIS'2013), University of uxembourg
(Presentation). Retrieved from https://publikationen.bibliothek.kit.edu/1000035777
Voß, J., Ledl, A., Balakrishnan, U. (2016). Uniform description and access to
Knowledge Organization Systems with BARTOC and JSKOS. Proceedings of TOTh
conference 2016. Retrieved from https://zenodo.org/record/438019
Voß, J. (2016). JSKOS data format for knowledge organization systems. Retrieved
from https://gbv.github.io/jskos/jskos.html
Walter, A.-K., Mayr, P., Petras, V., Baerisch, S. (2007). Kompetenzzentrum
Modellbildung und Heterogenitätsbehandlung (KoMoHe). Retrieved from
https://www.gesis.org/forschung/drittmittelprojekte/archiv/komohe/