Reproducible knowledge graph construction: a mapping-based wrapper over KGC engines

Hoogeveen, Dim C.M. (2021). Reproducible knowledge graph construction: a mapping-based wrapper over KGC engines. Thesis (Master thesis), E.T.S. de Ingenieros Informáticos (UPM).

Description

Title: Reproducible knowledge graph construction: a mapping-based wrapper over KGC engines
Author/s:
  • Hoogeveen, Dim C.M.
Contributor/s:
  • Corcho García, Óscar
  • Chaves Fraga, David
Item Type: Thesis (Master thesis)
Masters title: Data Science
Date: 4 July 2021
Subjects:
Faculty: E.T.S. de Ingenieros Informáticos (UPM)
Department: Inteligencia Artificial
Creative Commons Licenses: Recognition - No derivative works - Non commercial

Full text

[img]
Preview
PDF - Requires a PDF viewer, such as GSview, Xpdf or Adobe Acrobat Reader
Download (36MB) | Preview

Abstract

Knowledge graphs have shown to be effective at capturing domain knowledge to extract value from data sources. Multiple knowledge graph construction engines have emerged to create knowledge graphs; each with its own use-case, configuration settings, accepted mapping language(s) and implementation language. This variety of tools also has its downsides in that it hinders reproducibility; a mapping could lead to different results depending on the engine used. Furthermore, switching between engines is not a trivial task, since all configuration settings have to be set for each engine. In this thesis, an abstraction layer is proposed over the knowledge graph construction engines to tackle this. The abstraction layer consists of a newly proposed extension, that is mapping language independent, which declares the configurations; each configuration follows from an investigation into the implicit settings of the creation of a knowledge graph. The extension, called Templated Mapping Language (TML), is combined with a parser that generates a mapping and the configuration file and/or command for a chosen engine. An implementation is built to show the viability of the proposal, which is hosted as a website for easy access and usage for a wide audience. The work presents a starting point to make knowledge graph construction more accessible by abstracting away the intricacies in setting up an engine.

More information

Item ID: 68806
DC Identifier: https://oa.upm.es/68806/
OAI Identifier: oai:oa.upm.es:68806
Deposited by: Biblioteca Facultad de Informatica
Deposited on: 13 Oct 2021 09:10
Last Modified: 13 Oct 2021 09:10
  • Logo InvestigaM (UPM)
  • Logo GEOUP4
  • Logo Open Access
  • Open Access
  • Logo Sherpa/Romeo
    Check whether the anglo-saxon journal in which you have published an article allows you to also publish it under open access.
  • Logo Dulcinea
    Check whether the spanish journal in which you have published an article allows you to also publish it under open access.
  • Logo de Recolecta
  • Logo del Observatorio I+D+i UPM
  • Logo de OpenCourseWare UPM