Automated Information Retrieval for RDF-star-based Knowledge Graph Construction from Texts

textToRDFGraph : Automated Information Retrieval for RDF-star-based Knowledge Graph Construction from Texts This project aims to develop an automatic information retrieval pipeline based on Natural Language Processing and Semantic Web technologies to construct a knowledge graph from textual data. It explores how we can extract named entities from texts to create linked open data networks and then use the open network data to extract more text information and unify the resources representing identical entities retrieved from texts in different languages. The primary purpose of this project is thus to construct a graph from textual information that can be queried for the embedded references in texts regardless of the text language. A high-performing query requires the data to be stored optimally with all the metadata about extracted statements necessary for a faithful citation and reliable representation of data and metadata. Using RDF-star technology in graph construction carries this objective; therefore, this project shows how to use this technology to comprehensively represent textual data and facilitate complex queries required for humanities research. As a proof of concept and to implement and test the pipeline based on actual data, this project also generates an RDF-star-based digital edition of Jacob Bernoulli's historical travel journal, Reisebüchlein, using this pipeline. Nachwuchsförderung Forschungfonds of the University of Basel funded the project.