This talk describes the process of creating a digital visualization tool, based on Hugh Olmstead’s plectogram, to help the critical editor (Romanchuk) recover the original shape of the first episodes of the epic Digenis Akritis from the “reshuffled” abridgment, and identify variant readings from the Amazonian “cento” or patchwork quilt and incorporate them in their original places; and to help readers follow the critical reconstruction and understand pre-modern editing practices more generally.
In this talk, Dr. Keydar will present the results of a new study analyzing large testimonial dataset using unsupervised topic modeling. Applying LDA topic modeling to a corpus of court transcripts taken from a case before the International Criminal Tribunal for Former Yugoslavia (ICTY), Dr. Keydar will present a novel method for empirically assessing the international tribunals’ capacity to listen to large numbers of eye-witnesses. Harnessing the large quantity of testimonies, she uses topic modeling in order to explore latent themes, semantic fields and gaps between the language of the victims and that of the court.
A computational pipeline is a way of modeling the flow of information through a sequence of programs or operations, so that the output of each step becomes the input to the next. The decomposition of complex operations into discrete individual steps, each of which does only one thing, and none of which depends on knowing how the others operate, supports distributed, modular development; enhances maintenance and sustainability; and enables reusability. In this workshop we will examine how the concept of computational pipelines can be extended to model the planning, development, and deployment of digital textual editions. Participants are encouraged to come prepared to discuss their own digital edition plans and projects.
2 talks on topic modeling projects in progress: Todd Presner's talk focuses on topic modeling a multilingual collection of 120 Holocaust survivor testimonies made in 1946 by David Boder (on a wire recorder in Displaced Persons Camps). Dave Shepard will share his topic model of authors writing in English in the first half of the seventeenth century to place Andrew Marvell's “Horatian Ode” in its broader context and to show that, far from being ambivalent, “Horatian Ode” expresses a subtle, but clear, critique of Cromwell.
Our next meeting will be Wednesday, June 5 between 1:00 and 2:00pm in the SIL. We will continue our conversation about word embeddings and experiment with their practical application in our next meeting. In the meantime, feel free to begin playing with the code on your own.
In this meeting, we will discuss Andrew Piper's Enumerations, with a particular focus on Chapter 5. This book presents Piper's arguments about several methods for distant reading. Chapter 5 in particular looks at distant reading of characterization and character relationships. Also, during this session, we will do some planning for the rest of the year...
This talk will describe how “Early Modern Women in Ottoman-Algeria” employs user-friendly tools, including Recogito, Gephi, and Tableau to recover the experience of marginalized peoples through text mining and network analysis.