Child pages
  • Towards Cultural-Scale Models of Full-Text project

Versions Compared


  • This line was added.
  • This line was removed.
  • Formatting was changed.


 Figure. Topic Alignments. Subcorpus models (k = {20, 40, 60, 80}) for all five selected LCCO subject headings. In general, coarser models (i.e., those with lower number of topics) achieve the worst-whole-corpus-performance with a smaller subcorpus size than fine-grained models (i.e., those with higher number of topics).

Community Impact

The Topic Explorer has become a core component of the HTRC Data Capsule. In addition to our research team’s presentations at the HTRC UnCamp 2015, a tutorial at JCDL 2015, and at the HathiTrust User’s Meeting, the work was presented by the larger HTRC community at the Humanities Intensive Learning and Teaching (HILT) conference in Indianapolis in July.


Jaimie Murdock, Jiaan Zeng, Colin Allen. Project report 

Jaimie Murdock, Jiaan Zeng, and Robert H McDonald. Topic Exploration with the HTRC Data Capsule for Non-Consumptive Research. In JCDL ’15 Proceedings of the 15th ACM/IEEE-CS joint conference on Digital libraries, Knoxville, Tennessee, USA, 2015. ACM Press.