Extracting and Modeling Historical Events to Enhance Searching and Browsing of Digital Cultural Heritage Collections

Currently, cultural heritage portals limit their users to search only for individual objects and not for objects related to some historical narrative. Typically, most musea select objects for an exhibition based on the story they want to tell the public, but in digital collections this context can currently not be made explicit as the historical context is not part of the object annotations. From previous experiences with cultural heritage portals such as Europeana1 and CHIP2, we observed that adding historical events to object descriptions is valuable for the grounding of cultural heritage objects in their historical context as events represent important changepoints in time and form the basic units of the historical narrative. Further, historical event descriptions are comprised of actors, locations and timestamps that are in some cases already present as facets of the object annotation. As such, adding events to object descriptions can enhance browsing and searching of cultural heritage collections as the events unify otherwise unrelated but historical relevant facets of object annotations. The problem motivating this research is threefold: (1) There is no standard practice in cultural heritage organizations to include events within the object annotation, e.g. there is no shared vocabulary for the (historical) event descriptions, neither for different historical terms and concepts. The creation of such vocabulary is important in order to ensure alignment between the descriptions of historical events, which typically vary over time. (2) A multitude of different historical perspectives and interpretations of the same historical event exist. As a result, a variety of expressions can be used to describe the same event which implies problems for creating thesauri and vocabularies. (3) There is a lack of a structured representation of historical events, which could provide meaningful relationships between the events and pertaining objects. This PhD research is situated in the context of two projects: (1) The Semantics of History project3 with focus on the modeling and extraction of events and their perspectives from historical text documents; (2)Agora4 with focus on