Towards a Semantics-Based Recommendation System for Cultural Heritage Collections

While the use of semantic technologies is now commonplace in the cultural heritage sector and several semantically annotated cultural heritage datasets are publicly available, there are few examples of cultural portals that exploit these datasets and technologies to improve the experience of visitors to their online collections. Aiming to address this gap, this paper explores methods for semantics-based recommendations aimed at visitors to cultural portals who want to explore online collections. The proposed methods exploit the rich semantic metadata in a cultural heritage dataset and the capabilities of a graph database system to improve the accuracy of searches through the collection and the quality of the recommendations provided to the user. The methods were developed and tested with the Archive of the Art Textbooks of Elementary and Public Schools in the Japanese Colonial Period. However, they can easily be adapted to any cultural heritage collection dataset modelled in RDF.