evidence
doc2vec-based assisted close reading with support for abstract concept-based search and context-based search
Ego Documents Events modelling – how individuals recall mass violence
Much of our historical knowledge is based on oral or written accounts of eyewitnesses, particularly in cases of war and violence, when regular ways of documentation and record keeping are often absent. EviDENce studies how eyewitnesses have reported on violence, and how this may have changed over time. We use a collection of nearly 500 oral history interview transcripts about the Second World War [1] as well as the ego-documents (diaries, memoires, letters, autobiographies) available in Nederlab [2], covering a time span of 5 centuries.
Whereas humanities scholars are good at assessing texts for their relevance in relation to a particular topic or research question such as this, automating this assessment process, for example for distant reading or creating large corpora, is known to be problematic, especially when it comes to implicit mentions. EviDENce compares existing NLP methods to detect fragments containing mentions of such an ambiguous concept as violence, in a way that meets the standards of historical research.
Team members:
Susan Hogervorst, Open Universiteit Nederland (PI)Marieke van Erp (KNAW Humanities Cluster)Hennie Brugman (KNAW Humanities Cluster)Jeroen Willemsen (Open Universiteit Nederland) Edwin Klijn (NIOD Netwerk Oorlogsbronnen)Meiert Grootes (Netherlands eScience Center)
[1] http://getuigenverhalen.nl/home. The collection is hosted by the NIOD Institute of War, Holocaust- and Genocide studies in Amsterdam, and stored at DANS Data Archiving and Networked Services, https://dans.knaw.nl/en/about.
[2]https://www.nederlab.nl/onderzoeksportaal/?action=verkennen
An interactive web-based platform to investigate the dynamics of global corporate networks
Word vector text mining change and continuity in conceptual history
A new approach to the history of parliamentary communication and discourse
Facilitating and supporting large-scale text mining in the field of digital humanities
doc2vec-based assisted close reading with support for abstract concept-based search and context-based search
A flexible solution to build text mining workflows that allows you to quickly combine Natural Language Processing tools from different sources.