Mapping Art to a Knowledge Graph: Using Data for Exploring the Relations among Visual Objects in Renaissance Art

Abstract

Graph-like structures, which are increasingly popular in data representation, stand out since they enable the integration of information from multiple sources. At the same time, clustering algorithms applied on graphs allow for group entities based on similar characteristics, and discover statistically important information. This paper aims to explore the associations between the visual objects of the Renaissance in the Europeana database, based on the results of topic modeling and analysis. For this purpose, we employ Europeana’s Search and Report API to investigate the relations between the visual objects from this era, spanning from the 14th to the 17th century, and to create clusters of similar art objects. This approach will lead in transforming a cultural heritage database with semantic technologies into a dynamic digital knowledge representation graph that will relate art objects and their attributes. Based on associations between metadata, we will conduct a statistic analysis utilizing the knowledge graph of Europeana and topic modeling analysis.

Keywords

• machine learning
• data mining
• visualization
• topic modeling
• cluster analysis
• knowledge graph