From Books to Knowledge Graphs
Authors:
Natallia Kokash,
Matteo Romanello,
Ernest Suyver,
Giovanni Colavizza
Abstract:
The digital transformation of the scientific publishing industry has led to dramatic improvements in content discoverability and information analytics. Unfortunately, these improvements have not been uniform across research areas. The scientific literature in the arts, humanities and social sciences (AHSS) still lags behind, in part due to the scale of analog backlogs, the persisting importance of…
▽ More
The digital transformation of the scientific publishing industry has led to dramatic improvements in content discoverability and information analytics. Unfortunately, these improvements have not been uniform across research areas. The scientific literature in the arts, humanities and social sciences (AHSS) still lags behind, in part due to the scale of analog backlogs, the persisting importance of national languages, and a publisher ecosystem made of many, small or medium enterprises. We propose a bottom-up approach to support publishers in creating and maintaining their own publication knowledge graphs in the open domain. We do so by releasing a pipeline able to extract structured information from the bibliographies and indexes of AHSS publications, disambiguate, normalize and export it as linked data. We test the proposed pipeline on Brill's Classics collection, and release an implementation in open source for further use and improvement.
△ Less
Submitted 10 March, 2023; v1 submitted 22 April, 2022;
originally announced April 2022.
