A Pipeline Supporting a Smart Access to Historical Documents based on a Rich Semantic Representation of Their Content: A Case Study on Time Expressions

This work is part of two ongoing projects whose main goal is to demonstrate how semantic technologies can support an effective access to historical archives. In this paper we present a full pipeline, from rough texts up to the final user interface, aimed at creating and exploiting such representations. The pipeline is structured in three modules - handling information extraction, semantic representations, and queries - and offers external applications the possibility of accessing, and thus re-using, the output of each module, by providing a tagged text, a SPARQL endpoint, and a RESTful web service. In the paper, we describe the details of a proof-of-concept implementation of the pipeline architecture that focuses on time expressions. Moreover, we present an example application that exploits the pipeline to enable users to access historical documents by searching and browsing events and time specifications, thus demonstrating the effectiveness of an access to historical texts based on a rich semantic representation of their content.