Predicting Quality of Crowdsourced Annotations Using Graph Kernels

Abstract : Annotations obtained by Cultural Heritage institutions from the crowd need to be automatically assessed for their quality. Machine learning using graph kernels is an effective technique to use structural information in datasets to make predictions. We employ the Weisfeiler-Lehman graph kernel for RDF to make predictions about the quality of crowdsourced annotations in Steve.museum dataset, which is modelled and enriched as RDF. Our results indicate that we could predict quality of crowdsourced annotations with an accuracy of 75 %. We also employ the kernel to understand which features from the RDF graph are relevant to make predictions about different categories of quality.
Type de document :
Communication dans un congrès
Christian Damsgaard Jensen; Stephen Marsh; Theo Dimitrakos; Yuko Murayama. 9th IFIP International Conference on Trust Management (TM), May 2015, Hamburg, Germany. IFIP Advances in Information and Communication Technology, AICT-454, pp.134-148, 2015, Trust Management IX. 〈10.1007/978-3-319-18491-3_10〉
Liste complète des métadonnées

Littérature citée [22 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-01416219
Contributeur : Hal Ifip <>
Soumis le : mercredi 14 décembre 2016 - 11:09:04
Dernière modification le : samedi 18 novembre 2017 - 18:16:02
Document(s) archivé(s) le : mercredi 15 mars 2017 - 13:04:19

Fichier

337890_1_En_10_Chapter.pdf
Fichiers produits par l'(les) auteur(s)

Licence


Distributed under a Creative Commons Paternité 4.0 International License

Identifiants

Citation

Archana Nottamkandath, Jasper Oosterman, Davide Ceolin, Gerben Vries, Wan Fokkink. Predicting Quality of Crowdsourced Annotations Using Graph Kernels. Christian Damsgaard Jensen; Stephen Marsh; Theo Dimitrakos; Yuko Murayama. 9th IFIP International Conference on Trust Management (TM), May 2015, Hamburg, Germany. IFIP Advances in Information and Communication Technology, AICT-454, pp.134-148, 2015, Trust Management IX. 〈10.1007/978-3-319-18491-3_10〉. 〈hal-01416219〉

Partager

Métriques

Consultations de la notice

60

Téléchargements de fichiers

10